Merge branch 'develop' of github.com:baidu/Paddle into feature/fast_ci

cmake/external/boost.cmake

@@ -15,9 +15,9 @@
 include(ExternalProject)
 
 set(BOOST_PROJECT       "extern_boost")
-set(BOOST_VER           "1.66.0")
-set(BOOST_TAR           "boost_1_66_0")
-set(BOOST_URL           "https://dl.bintray.com/boostorg/release/${BOOST_VER}/source/${BOOST_TAR}.tar.gz")
+set(BOOST_VER           "1.41.0")
+set(BOOST_TAR           "boost_1_41_0")
+set(BOOST_URL           "http://sourceforge.net/projects/boost/files/boost/${BOOST_VER}/${BOOST_TAR}.tar.gz")
 set(BOOST_SOURCES_DIR ${THIRD_PARTY_PATH}/boost)
 set(BOOST_DOWNLOAD_DIR  "${BOOST_SOURCES_DIR}/src/${BOOST_PROJECT}")
 set(BOOST_INCLUDE_DIR "${BOOST_DOWNLOAD_DIR}/${BOOST_TAR}" CACHE PATH "boost include directory." FORCE)

doc/api/v2/fluid/layers.rst

@@ -534,3 +534,8 @@ row_conv
 --------
 ..  autofunction:: paddle.v2.fluid.layers.row_conv
     :noindex:
+
+multiplex
+---------
+..  autofunction:: paddle.v2.fluid.layers.multiplex
+    :noindex:

doc/design/support_new_device.md

@@ -2,9 +2,9 @@
 
 ## Background
 
-Deep learning has a high demand for computing resources. New high-performance devices and computing libraries are appearing very frequently. Deep learning frameworks have to integrate these high-performance devices and computing libraries flexibly and efficiently.
+Deep learning has a high demand for computing resources. New high-performance devices and computing libraries are appearing very frequently. Deep learning frameworks have to integrate these high-performance devices and computing libraries in a flexible and efficient manner.
 
-On one hand, hardware and computing libraries usually do not have a one-to-one correspondence. For example,Intel CPUs support Eigen and MKL computing libraries while Nvidia GPUs support Eigen and cuDNN computing libraries. We have to implement operator specific kernels for each computing library.
+On one hand, hardware and computing libraries usually do not have a one-to-one correspondence. For example, Intel CPUs support Eigen and MKL computing libraries while Nvidia GPUs support Eigen and cuDNN computing libraries. We have to implement operator specific kernels for each computing library.
 
 On the other hand, users usually do not want to care about the low-level hardware and computing libraries when writing a neural network configuration. In Fluid, `Layer` is exposed in `Python`, and `Operator` is exposed in `C++`. Both `Layer` and `Operator` are hardware independent.
 
@@ -17,7 +17,7 @@ For a general overview of fluid, please refer to the [overview doc](https://gith
 
 There are mainly three parts that we have to consider while integrating a new device/library:
 
-- Place and DeviceContext: indicates the device id and manages hardware resources
+- Place and DeviceContext: indicate the device id and manage hardware resources
 
 - Memory and Tensor: malloc/free data on certain device
 
@@ -25,10 +25,10 @@ There are mainly three parts that we have to consider while integrating a new de
 
 ### Place and DeviceContext
 
-Please remind that device and computing library are not one-to-one corresponding. A device can have a lot of computing libraries and a computing library can also support several devices.
+Please note that device and computing library are not one-to-one corresponding. A device can have a lot of computing libraries and a computing library can also support several devices.
 
 #### Place
-Fluid uses class [Place](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/platform/place.h#L55) to represent the device memory where data is located. If we add another device, we have to add corresponding `DevicePlace`.
+Fluid uses class [Place](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/platform/place.h#L55) to represent the device memory where data is located. If we add another device, we have to add the corresponding `DevicePlace`.
 
 ```
         |   CPUPlace
@@ -144,7 +144,7 @@ class Tensor {
 };
 ```
 
-`Placeholder` is used to delay memory allocation; that is, we can first define a tensor, using `Resize` to configure its shape, and then call `mutuable_data` to allocate the actual memory.
+`Placeholder` is used to delay memory allocation; that is, we can first define a tensor, using `Resize` to configurate its shape, and then call `mutuable_data` to allocate the actual memory.
 
 ```cpp
 paddle::framework::Tensor t;
@@ -163,7 +163,7 @@ Fluid implements computing units based on different DeviceContexts. Some computi
 
 Let's take [MaxOutFunctor](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/operators/math/maxouting.h#L27) as an example:
 
-The interface is defined in header file.
+The interface is defined in the header file.
 
 ```
 template <typename DeviceContext, typename T>
@@ -203,7 +203,7 @@ class MaxOutFunctor<platform::CUDADeviceContext, T> {
 ```
 
 
-We get computing handle from a concrete DeviceContext, and make compution on tensors.
+We first obtain the computing handle from a concrete DeviceContext, and then compute on tensors.
 
 The implemention of `OpKernel` is similar to math functors, the extra thing we need to do is to register the OpKernel in a global map.
 
@@ -231,7 +231,7 @@ REGISTER_OP_CUDA_KERNEL(
 
 ## Advanced topics: How to switch between different Device/Library
 
-Generally, we will impelement OpKernel for all Device/Library of an Operator. We can easily train a Convolutional Neural Network in GPU. However, some OpKernel is not sutibale on a specific Device. For example, crf operator can only run on CPU, whereas most other operators can run at GPU. To achieve high performance in such circumstance, we have to switch between different Device/Library.
+Generally, we will implement OpKernel for all Device/Library of an Operator. We can easily train a Convolutional Neural Network in GPU. However, some OpKernel is not suitable on a specific Device. For example, crf operator can only run on CPU, whereas most other operators can run on GPU. To achieve high performance in such circumstance, we have to switch between different Device/Library.
 
 
 For more details, please refer to following docs:

doc/getstarted/build_and_install/pip_install_cn.rst

@@ -39,6 +39,7 @@ PaddlePaddle可以使用常用的Python包管理工具
 
     "cpu_avx_mkl", "`paddlepaddle-0.11.0-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxCp27cp27mu/.lastSuccessful/paddlepaddle-0.11.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle-0.11.0-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxCp27cp27mu/.lastSuccessful/paddlepaddle-0.11.0-cp27-cp27m-linux_x86_64.whl>`_", "`paddle.tgz <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxCp27cp27mu/.lastSuccessful/paddle.tgz>`_"
     "cpu_avx_openblas", "`paddlepaddle-0.11.0-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxOpenblas/.lastSuccessful/paddlepaddle-0.11.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle-0.11.0-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxOpenblas/.lastSuccessful/paddlepaddle-0.11.0-cp27-cp27m-linux_x86_64.whl>`_", "暂无"
+    "cpu_noavx_openblas", "`paddlepaddle-0.11.0-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuNoavxOpenblas/.lastSuccessful/paddlepaddle-0.11.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle-0.11.0-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuNoavxOpenblas/.lastSuccessful/paddlepaddle-0.11.0-cp27-cp27m-linux_x86_64.whl>`_", "暂无"
     "cuda7.5_cudnn5_avx_mkl", "`paddlepaddle_gpu-0.11.0-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda75cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-0.11.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle_gpu-0.11.0-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda75cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-0.11.0-cp27-cp27m-linux_x86_64.whl>`_", "`paddle.tgz <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda75cudnn5cp27cp27mu/.lastSuccessful/paddle.tgz>`_"
     "cuda8.0_cudnn5_avx_mkl", "`paddlepaddle_gpu-0.11.0-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda80cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-0.11.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle_gpu-0.11.0-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda80cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-0.11.0-cp27-cp27m-linux_x86_64.whl>`_", "`paddle.tgz <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda80cudnn5cp27cp27mu/.lastSuccessful/paddle.tgz>`_"
     "cuda8.0_cudnn7_avx_mkl", "`paddlepaddle_gpu-0.11.0-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda8cudnn7cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-0.11.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle_gpu-0.11.0-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda8cudnn7cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-0.11.0-cp27-cp27m-linux_x86_64.whl>`_", "`paddle.tgz <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda8cudnn7cp27cp27mu/.lastSuccessful/paddle.tgz>`_"

doc/getstarted/build_and_install/pip_install_en.rst

@@ -42,6 +42,7 @@ If the links below shows up the login form, just click "Log in as guest" to star
 
     "cpu_avx_mkl", "`paddlepaddle-0.11.0-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxCp27cp27mu/.lastSuccessful/paddlepaddle-0.11.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle-0.11.0-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxCp27cp27mu/.lastSuccessful/paddlepaddle-0.11.0-cp27-cp27m-linux_x86_64.whl>`_", "`paddle.tgz <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxCp27cp27mu/.lastSuccessful/paddle.tgz>`_"
     "cpu_avx_openblas", "`paddlepaddle-0.11.0-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxOpenblas/.lastSuccessful/paddlepaddle-0.11.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle-0.11.0-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxOpenblas/.lastSuccessful/paddlepaddle-0.11.0-cp27-cp27m-linux_x86_64.whl>`_", "Not Available"
+    "cpu_noavx_openblas", "`paddlepaddle-0.11.0-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuNoavxOpenblas/.lastSuccessful/paddlepaddle-0.11.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle-0.11.0-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuNoavxOpenblas/.lastSuccessful/paddlepaddle-0.11.0-cp27-cp27m-linux_x86_64.whl>`_", "Not Available"
     "cuda7.5_cudnn5_avx_mkl", "`paddlepaddle_gpu-0.11.0-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda75cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-0.11.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle_gpu-0.11.0-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda75cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-0.11.0-cp27-cp27m-linux_x86_64.whl>`_", "`paddle.tgz <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda75cudnn5cp27cp27mu/.lastSuccessful/paddle.tgz>`_"
     "cuda8.0_cudnn5_avx_mkl", "`paddlepaddle_gpu-0.11.0-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda80cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-0.11.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle_gpu-0.11.0-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda80cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-0.11.0-cp27-cp27m-linux_x86_64.whl>`_", "`paddle.tgz <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda80cudnn5cp27cp27mu/.lastSuccessful/paddle.tgz>`_"
     "cuda8.0_cudnn7_avx_mkl", "`paddlepaddle_gpu-0.11.0-cp27-cp27mu-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda8cudnn7cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-0.11.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle_gpu-0.11.0-cp27-cp27m-linux_x86_64.whl <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda8cudnn7cp27cp27mu/.lastSuccessful/paddlepaddle_gpu-0.11.0-cp27-cp27m-linux_x86_64.whl>`_", "`paddle.tgz <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda8cudnn7cp27cp27mu/.lastSuccessful/paddle.tgz>`_"

paddle/framework/CMakeLists.txt

@@ -74,8 +74,10 @@ cc_library(backward SRCS backward.cc DEPS net_op)
 cc_test(backward_test SRCS backward_test.cc DEPS backward recurrent_op device_context fill_constant_op)
 cc_library(lod_rank_table SRCS lod_rank_table.cc DEPS lod_tensor)
 
+cc_library(feed_fetch_method SRCS feed_fetch_method.cc DEPS lod_tensor scope glog)
+
 cc_library(executor SRCS executor.cc DEPS op_registry device_context scope
-framework_proto backward glog lod_rank_table profiler)
+framework_proto backward glog lod_rank_table profiler feed_fetch_method)
 
 cc_library(prune SRCS prune.cc DEPS framework_proto)
 cc_test(prune_test SRCS prune_test.cc DEPS op_info prune recurrent_op device_context)

paddle/framework/channel.h

+/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License. */
+
+#pragma once
+#include <condition_variable>
+#include <mutex>
+#include <queue>
+
+namespace paddle {
+namespace framework {
+
+template <typename T>
+class Channel {
+ public:
+  explicit Channel(std::size_t capacity) : capacity_(capacity) {}
+
+  void Send(T* channel_element) {
+    std::unique_lock<std::mutex> lock(mu_);
+
+    if (IsBounded()) {
+      full_cond_var_.wait(lock, [this]() {
+        bool capacity_valid = capacity_ > 0 ? !IsCapacityFull() : true;
+        return capacity_valid;
+      });
+    }
+    channel_.push_back(std::move(*channel_element));
+
+    lock.unlock();
+    empty_cond_var_.notify_one();
+  }
+
+  T* Receive() {
+    std::unique_lock<std::mutex> lock(mu_);
+    empty_cond_var_.wait(lock, [this]() { return !channel_.empty(); });
+
+    T* channel_element = std::move(channel_.front());
+    channel_.pop_front();
+
+    NotifyAllSenders(&lock);
+    return channel_element;
+  }
+
+  size_t Size() {
+    std::unique_lock<std::mutex> lock(mu_);
+    return channel_.size();
+  }
+
+  void Clear() {
+    std::unique_lock<std::mutex> lock(mu_);
+    channel_.clear();
+
+    NotifyAllSenders(&lock);
+  }
+
+ private:
+  std::size_t capacity_;
+  std::mutex mu_;
+  std::condition_variable empty_cond_var_;
+  std::condition_variable full_cond_var_;
+  std::deque<T> channel_;
+
+ private:
+  void NotifyAllSenders(std::unique_lock<std::mutex>* lock) {
+    if (IsBounded()) {
+      lock->unlock();
+      full_cond_var_.notify_one();
+    }
+  }
+
+  bool IsBounded() const { return capacity_ > 0; }
+
+  bool IsCapacityFull() const { return channel_.size() >= capacity_; }
+};
+
+}  // namespace operator
+}  // namespace paddle

paddle/framework/data_type.h

@@ -79,5 +79,33 @@ inline void VisitDataType(proto::DataType type, Visitor visitor) {
   }
 }
 
+inline std::string DataTypeToString(const proto::DataType type) {
+  using namespace paddle::framework::proto;
+  switch (type) {
+    case DataType::FP16:
+      return "float16";
+    case DataType::FP32:
+      return "float32";
+    case DataType::FP64:
+      return "float64";
+    case DataType::INT16:
+      return "int16";
+    case DataType::INT32:
+      return "int32";
+    case DataType::INT64:
+      return "int64";
+    case DataType::BOOL:
+      return "bool";
+    default:
+      PADDLE_THROW("Not support type %d", type);
+  }
+}
+
+inline std::ostream& operator<<(std::ostream& out,
+                                const proto::DataType& type) {
+  out << DataTypeToString(type);
+  return out;
+}
+
 }  // namespace framework
 }  // namespace paddle

paddle/framework/executor.cc

@@ -17,6 +17,7 @@ limitations under the License. */
 #include <set>
 
 #include "gflags/gflags.h"
+#include "paddle/framework/feed_fetch_method.h"
 #include "paddle/framework/feed_fetch_type.h"
 #include "paddle/framework/lod_rank_table.h"
 #include "paddle/framework/lod_tensor_array.h"
@@ -149,5 +150,164 @@ void Executor::Run(const ProgramDesc& pdesc, Scope* scope, int block_id,
   }
 }
 
+// Check whether the block already has feed operators and feed_holder.
+// Return false if the block does not have any feed operators.
+// If some feed operators have been prepended to the block, check that
+// the info contained in these feed operators matches the feed_targets
+// and feed_holder_name. Raise exception when any mismatch is found.
+// Return true if the block has feed operators and holder of matching info.
+static bool has_feed_operators(
+    BlockDesc* block, std::map<std::string, const LoDTensor*>& feed_targets,
+    const std::string& feed_holder_name) {
+  size_t feed_count = 0;
+  for (auto* op : block->AllOps()) {
+    if (op->Type() == kFeedOpType) {
+      feed_count++;
+      PADDLE_ENFORCE_EQ(op->Input("X")[0], feed_holder_name,
+                        "Input to feed op should be '%s'", feed_holder_name);
+      std::string feed_target_name = op->Output("Out")[0];
+      PADDLE_ENFORCE(
+          feed_targets.find(feed_target_name) != feed_targets.end(),
+          "Feed operator output name '%s' cannot be found in 'feed_targets'",
+          feed_target_name);
+    }
+  }
+
+  if (feed_count > 0) {
+    PADDLE_ENFORCE_EQ(
+        feed_count, feed_targets.size(),
+        "The number of feed operators should match 'feed_targets'");
+
+    // When feed operator are present, so should be feed_holder
+    auto var = block->FindVar(feed_holder_name);
+    PADDLE_ENFORCE_NOT_NULL(var, "Block should already have a '%s' variable",
+                            feed_holder_name);
+    PADDLE_ENFORCE_EQ(var->GetType(), proto::VarDesc::FEED_MINIBATCH,
+                      "'%s' variable should be 'FEED_MINIBATCH' type",
+                      feed_holder_name);
+  }
+
+  return feed_count > 0;
+}
+
+// Check whether the block already has fetch operators and fetch_holder.
+// Return false if the block does not have any fetch operators.
+// If some fetch operators have been appended to the block, check that
+// the info contained in these fetch operators matches the fetch_targets
+// and fetch_holder_name. Raise exception when any mismatch is found.
+// Return true if the block has fetch operators and holder of matching info.
+static bool has_fetch_operators(
+    BlockDesc* block, std::map<std::string, LoDTensor*>& fetch_targets,
+    const std::string& fetch_holder_name) {
+  size_t fetch_count = 0;
+  for (auto* op : block->AllOps()) {
+    if (op->Type() == kFetchOpType) {
+      fetch_count++;
+      PADDLE_ENFORCE_EQ(op->Output("Out")[0], fetch_holder_name,
+                        "Output of fetch op should be '%s'", fetch_holder_name);
+      std::string fetch_target_name = op->Input("X")[0];
+      PADDLE_ENFORCE(
+          fetch_targets.find(fetch_target_name) != fetch_targets.end(),
+          "Fetch operator input name '%s' cannot be found in 'fetch_targets'",
+          fetch_target_name);
+    }
+  }
+
+  if (fetch_count > 0) {
+    PADDLE_ENFORCE_EQ(
+        fetch_count, fetch_targets.size(),
+        "The number of fetch operators should match 'fetch_targets'");
+
+    // When fetch operator are present, so should be fetch_holder
+    auto var = block->FindVar(fetch_holder_name);
+    PADDLE_ENFORCE_NOT_NULL(var, "Block should already have a '%s' variable",
+                            fetch_holder_name);
+    PADDLE_ENFORCE_EQ(var->GetType(), proto::VarDesc::FETCH_LIST,
+                      "'%s' variable should be 'FETCH_LIST' type",
+                      fetch_holder_name);
+  }
+
+  return fetch_count > 0;
+}
+
+void Executor::Run(const ProgramDesc& program, Scope* scope,
+                   std::map<std::string, const LoDTensor*>& feed_targets,
+                   std::map<std::string, LoDTensor*>& fetch_targets,
+                   const std::string& feed_holder_name,
+                   const std::string& fetch_holder_name) {
+  auto* copy_program = new ProgramDesc(program);
+  auto* global_block = copy_program->MutableBlock(0);
+
+  if (!has_feed_operators(global_block, feed_targets, feed_holder_name)) {
+    // create feed_holder variable
+    auto* feed_holder = global_block->Var(feed_holder_name);
+    feed_holder->SetType(proto::VarDesc::FEED_MINIBATCH);
+    feed_holder->SetPersistable(true);
+
+    int i = 0;
+    for (auto& feed_target : feed_targets) {
+      std::string var_name = feed_target.first;
+      VLOG(3) << "feed target's name: " << var_name;
+
+      // prepend feed op
+      auto* op = global_block->PrependOp();
+      op->SetType(kFeedOpType);
+      op->SetInput("X", {feed_holder_name});
+      op->SetOutput("Out", {var_name});
+      op->SetAttr("col", {static_cast<int>(i)});
+      op->CheckAttrs();
+
+      i++;
+    }
+  }
+
+  // map the data of feed_targets to feed_holder
+  for (auto* op : global_block->AllOps()) {
+    if (op->Type() == kFeedOpType) {
+      std::string feed_target_name = op->Output("Out")[0];
+      int idx = boost::get<int>(op->GetAttr("col"));
+      SetFeedVariable(scope, *feed_targets[feed_target_name], feed_holder_name,
+                      idx);
+    }
+  }
+
+  if (!has_fetch_operators(global_block, fetch_targets, fetch_holder_name)) {
+    // create fetch_holder variable
+    auto* fetch_holder = global_block->Var(fetch_holder_name);
+    fetch_holder->SetType(proto::VarDesc::FETCH_LIST);
+    fetch_holder->SetPersistable(true);
+
+    int i = 0;
+    for (auto& fetch_target : fetch_targets) {
+      std::string var_name = fetch_target.first;
+      VLOG(3) << "fetch target's name: " << var_name;
+
+      // append fetch op
+      auto* op = global_block->AppendOp();
+      op->SetType(kFetchOpType);
+      op->SetInput("X", {var_name});
+      op->SetOutput("Out", {fetch_holder_name});
+      op->SetAttr("col", {static_cast<int>(i)});
+      op->CheckAttrs();
+
+      i++;
+    }
+  }
+
+  Run(*copy_program, scope, 0, true, true);
+
+  // obtain the data of fetch_targets from fetch_holder
+  for (auto* op : global_block->AllOps()) {
+    if (op->Type() == kFetchOpType) {
+      std::string fetch_target_name = op->Input("X")[0];
+      int idx = boost::get<int>(op->GetAttr("col"));
+      *fetch_targets[fetch_target_name] =
+          GetFetchVariable(*scope, fetch_holder_name, idx);
+    }
+  }
+
+  delete copy_program;
+}
+
 }  // namespace framework
 }  // namespace paddle

paddle/framework/executor.h

@@ -41,6 +41,12 @@ class Executor {
   void Run(const ProgramDesc&, Scope*, int, bool create_local_scope = true,
            bool create_vars = true);
 
+  void Run(const ProgramDesc& program, Scope* scope,
+           std::map<std::string, const LoDTensor*>& feed_targets,
+           std::map<std::string, LoDTensor*>& fetch_targets,
+           const std::string& feed_holder_name = "feed",
+           const std::string& fetch_holder_name = "fetch");
+
  private:
   const platform::Place place_;
 };

paddle/framework/feed_fetch_method.cc

+/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License. */
+
+#include "paddle/framework/feed_fetch_method.h"
+#include "glog/logging.h"
+#include "paddle/framework/variable.h"
+
+namespace paddle {
+namespace framework {
+
+void SetFeedVariable(Scope* scope, const LoDTensor& input,
+                     const std::string& var_name, size_t index) {
+  // If var_name Variable is not found in GlobalScope, a new variable will
+  // be created.
+  VLOG(3) << "SetFeedVariable name=" << var_name << " index=" << index;
+  Variable* g_feed_value = scope->Var(var_name);
+  auto& feed_inputs =
+      *(g_feed_value->GetMutable<std::vector<paddle::framework::LoDTensor>>());
+  if (index >= feed_inputs.size()) {
+    feed_inputs.resize(index + 1);
+  }
+  // shared data with input tensor
+  feed_inputs[index].ShareDataWith(input);
+  // set lod
+  feed_inputs[index].set_lod(input.lod());
+}
+
+LoDTensor& GetFetchVariable(const Scope& scope, const std::string& var_name,
+                            size_t index) {
+  // Since we want to fetch LodTensor from a variable, the variable must
+  // be created alreadly.
+  Variable* g_fetch_value = scope.FindVar(var_name);
+  PADDLE_ENFORCE(g_fetch_value->IsType<FeedFetchList>(),
+                 "Only %s can be invoked by GetFetchVariable",
+                 typeid(FeedFetchList).name());
+  auto& fetch_outputs = *g_fetch_value->GetMutable<FeedFetchList>();
+  auto& tensor = fetch_outputs[index];
+  VLOG(3) << "Fetch " << var_name << " with index " << index
+          << " shape= " << tensor.dims();
+  PADDLE_ENFORCE_LT(index, fetch_outputs.size());
+  return tensor;
+}
+
+}  // namespace framework
+}  // namespace paddle

paddle/framework/feed_fetch_method.h

@@ -13,46 +13,18 @@ See the License for the specific language governing permissions and
 limitations under the License. */
 
 #pragma once
-#include "glog/logging.h"
+
 #include "paddle/framework/feed_fetch_type.h"
 #include "paddle/framework/scope.h"
-#include "paddle/framework/variable.h"
 
 namespace paddle {
 namespace framework {
 
 void SetFeedVariable(Scope* scope, const LoDTensor& input,
-                     const std::string& var_name, size_t index) {
-  // If var_name Variable is not found in GlobalScope, a new variable will
-  // be created.
-  VLOG(3) << "SetFeedVariable name=" << var_name << " index=" << index;
-  Variable* g_feed_value = scope->Var(var_name);
-  auto& feed_inputs =
-      *(g_feed_value->GetMutable<std::vector<paddle::framework::LoDTensor>>());
-  if (index >= feed_inputs.size()) {
-    feed_inputs.resize(index + 1);
-  }
-  // shared data with input tensor
-  feed_inputs[index].ShareDataWith(input);
-  // set lod
-  feed_inputs[index].set_lod(input.lod());
-}
+                     const std::string& var_name, size_t index);
 
 LoDTensor& GetFetchVariable(const Scope& scope, const std::string& var_name,
-                            size_t index) {
-  // Since we want to fetch LodTensor from a variable, the variable must
-  // be created alreadly.
-  Variable* g_fetch_value = scope.FindVar(var_name);
-  PADDLE_ENFORCE(g_fetch_value->IsType<FeedFetchList>(),
-                 "Only %s can be invoked by GetFetchVariable",
-                 typeid(FeedFetchList).name());
-  auto& fetch_outputs = *g_fetch_value->GetMutable<FeedFetchList>();
-  auto& tensor = fetch_outputs[index];
-  VLOG(3) << "Fetch " << var_name << " with index " << index
-          << " shape= " << tensor.dims();
-  PADDLE_ENFORCE_LT(index, fetch_outputs.size());
-  return tensor;
-}
+                            size_t index);
 
 }  // namespace framework
 }  // namespace paddle

paddle/framework/op_kernel_type_test.cc

@@ -26,9 +26,9 @@ TEST(OpKernelType, ToString) {
   OpKernelType op_kernel_type(DataType::FP32, CPUPlace(), DataLayout::kNCHW,
                               LibraryType::kCUDNN);
 
-  ASSERT_EQ(
-      paddle::framework::KernelTypeToString(op_kernel_type),
-      "data_type[5]:data_layout[NCHW]:place[CPUPlace]:library_type[CUDNN]");
+  ASSERT_EQ(paddle::framework::KernelTypeToString(op_kernel_type),
+            "data_type[float32]:data_layout[NCHW]:place[CPUPlace]:library_type["
+            "CUDNN]");
 }
 
 TEST(OpKernelType, Hash) {

paddle/framework/threadpool.cc

 /* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve.
 
-Licensed under the Apache License, Version 2.0 (the "License");
-you may not use this file except in compliance with the License.
-You may obtain a copy of the License at
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
 
-    http://www.apache.org/licenses/LICENSE-2.0
+   http://www.apache.org/licenses/LICENSE-2.0
 
-Unless required by applicable law or agreed to in writing, software
-distributed under the License is distributed on an "AS IS" BASIS,
-WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-See the License for the specific language governing permissions and
-limitations under the License. */
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License. */
 
 #include "paddle/framework/threadpool.h"
 
 namespace paddle {
 namespace framework {
 
-std::unique_ptr<ThreadPool> ThreadPool::threadpool(nullptr);
-std::once_flag ThreadPool::init_flag;
+std::unique_ptr<ThreadPool> ThreadPool::threadpool_(nullptr);
+std::once_flag ThreadPool::init_flag_;
+
+ThreadPool* ThreadPool::GetInstance() {
+  std::call_once(init_flag_, &ThreadPool::Init);
+  return threadpool_.get();
+}
+
+void ThreadPool::Init() {
+  if (threadpool_.get() == nullptr) {
+    // TODO(Yancey1989): specify the max threads number
+    int num_threads = std::thread::hardware_concurrency();
+    PADDLE_ENFORCE_GT(num_threads, 0);
+    threadpool_.reset(new ThreadPool(num_threads));
+  }
+}
+
+ThreadPool::ThreadPool(int num_threads)
+    : total_threads_(num_threads), idle_threads_(num_threads), running_(true) {
+  threads_.resize(num_threads);
+  for (auto& thread : threads_) {
+    // TODO(Yancey1989): binding the thread on the specify CPU number
+    thread.reset(new std::thread(std::bind(&ThreadPool::TaskLoop, this)));
+  }
+}
+
+ThreadPool::~ThreadPool() {
+  {
+    // notify all threads to stop running
+    running_ = false;
+    scheduled_.notify_all();
+  }
+
+  for (auto& t : threads_) {
+    t->join();
+    t.reset(nullptr);
+  }
+}
+
+void ThreadPool::Wait() {
+  std::unique_lock<std::mutex> lock(mutex_);
+  completed_.wait(lock, [=] { return Done() == true; });
+}
+
+void ThreadPool::TaskLoop() {
+  while (running_) {
+    std::unique_lock<std::mutex> lock(mutex_);
+    scheduled_.wait(lock, [=] { return !tasks_.empty() || !running_; });
+
+    if (!running_) {
+      break;
+    }
+    // pop a task from the task queue
+    auto task = std::move(tasks_.front());
+    tasks_.pop();
+
+    --idle_threads_;
+    lock.unlock();
+
+    // run the task
+    task();
+
+    {
+      std::unique_lock<std::mutex> lock(mutex_);
+      ++idle_threads_;
+      if (Done()) {
+        completed_.notify_all();
+      }
+    }
+  }
+}
 
 }  // namespace framework
 }  // namespace paddle

paddle/framework/threadpool.h

@@ -20,52 +20,36 @@ limitations under the License. */
 #include <mutex>
 #include <queue>
 #include <thread>
+#include <vector>
 
 #include "paddle/platform/enforce.h"
 
 namespace paddle {
 namespace framework {
 
+// ThreadPool maintains a queue of tasks, and runs them using a fixed
+// number of threads.
 class ThreadPool {
  public:
   typedef std::packaged_task<void()> Task;
 
-  /**
-   * @brief   Get a instance of threadpool, the thread number will
-   *          be specified as the number of hardware thread contexts
-   */
-  static ThreadPool* GetInstance() {
-    std::call_once(init_flag, &ThreadPool::Init);
-    return threadpool.get();
-  }
+  // Returns the singleton of ThreadPool.
+  static ThreadPool* GetInstance();
 
-  ~ThreadPool() {
-    {
-      // notify all threads to stop running
-      running_ = false;
-      scheduled_.notify_all();
-    }
-
-    for (auto& t : threads_) {
-      t->join();
-      t.reset(nullptr);
-    }
-  }
+  ~ThreadPool();
 
-  int GetNumThreads() const { return num_threads_; }
+  // Returns the number of threads created by the constructor.
+  size_t Threads() const { return total_threads_; }
 
-  int GetAvailable() {
+  // Returns the number of currently idle threads.
+  size_t IdleThreads() {
     std::unique_lock<std::mutex> lock(mutex_);
-    return available_;
+    return idle_threads_;
   }
 
-  /**
-   * @brief   Push a function to the queue, and will be scheduled and
-   *          executed if a thread is available.
-   * @param[in] Task, will be pushed to the task queue.
-   * @return    std::future<void>, we could wait for the task finished by
-   *            f.wait().
-   */
+  // Run pushes a function to the task queue and returns a std::future
+  // object.  To wait for the completion of the task, call
+  // std::future::wait().
   template <typename Callback>
   std::future<void> Run(Callback fn) {
     std::unique_lock<std::mutex> lock(mutex_);
@@ -77,84 +61,40 @@ class ThreadPool {
     return f;
   }
 
-  /**
-   * @brief   Wait until all the tasks are completed.
-   */
-  void Wait() {
-    std::unique_lock<std::mutex> lock(mutex_);
-    completed_.wait(lock, [=] { return Done() == true; });
-  }
+  // Wait until all the tasks are completed.
+  void Wait();
 
  private:
   DISABLE_COPY_AND_ASSIGN(ThreadPool);
 
-  explicit ThreadPool(int num_threads)
-      : num_threads_(num_threads), available_(num_threads), running_(true) {
-    threads_.resize(num_threads);
-    for (auto& thread : threads_) {
-      // TODO(Yancey1989): binding the thread on the specify CPU number
-      thread.reset(new std::thread(std::bind(&ThreadPool::TaskLoop, this)));
-    }
-  }
+  explicit ThreadPool(int num_threads);
 
-  /**
-   * @brief   If the task queue is empty and avaialbe
-   *          is equal to the number of threads, means that
-   *          all tasks are completed.
-   *
-   *          Note: this function is not thread-safe.
-   *
-   * @return true if all tasks are completed.
-   */
-  bool Done() { return tasks_.empty() && available_ == num_threads_; }
-
-  void TaskLoop() {
-    while (running_) {
-      std::unique_lock<std::mutex> lock(mutex_);
-      scheduled_.wait(lock, [=] { return !tasks_.empty() || !running_; });
-
-      if (!running_) {
-        break;
-      }
-      // pop a task from the task queue
-      auto task = std::move(tasks_.front());
-      tasks_.pop();
-
-      --available_;
-      lock.unlock();
-
-      // run the task
-      task();
-
-      {
-        std::unique_lock<std::mutex> lock(mutex_);
-        ++available_;
-        if (Done()) {
-          completed_.notify_all();
-        }
-      }
-    }
-  }
+  // If the task queue is empty and avaialbe is equal to the number of
+  // threads, means that all tasks are completed.  Note: this function
+  // is not thread-safe.  Returns true if all tasks are completed.
+  // Note: don't delete the data member total_threads_ and use
+  // threads_.size() instead; because you'd need to lock the mutex
+  // before accessing threads_.
+  bool Done() { return tasks_.empty() && idle_threads_ == total_threads_; }
 
-  static void Init() {
-    if (threadpool.get() == nullptr) {
-      // TODO(Yancey1989): specify the max threads number
-      int num_threads = std::thread::hardware_concurrency();
-      PADDLE_ENFORCE_GT(num_threads, 0);
-      threadpool.reset(new ThreadPool(num_threads));
-    }
-  }
+  // The constructor starts threads to run TaskLoop, which retrieves
+  // and runs tasks from the queue.
+  void TaskLoop();
+
+  // Init is called by GetInstance.
+  static void Init();
 
  private:
-  static std::unique_ptr<ThreadPool> threadpool;
-  static std::once_flag init_flag;
+  static std::unique_ptr<ThreadPool> threadpool_;
+  static std::once_flag init_flag_;
 
-  int num_threads_;
-  int available_;
-  bool running_;
-  std::queue<Task> tasks_;
   std::vector<std::unique_ptr<std::thread>> threads_;
+  const size_t total_threads_;
+  size_t idle_threads_;
+
+  std::queue<Task> tasks_;
   std::mutex mutex_;
+  bool running_;
   std::condition_variable scheduled_;
   std::condition_variable completed_;
 };

paddle/framework/threadpool_test.cc

@@ -22,11 +22,7 @@ namespace framework = paddle::framework;
 void do_sum(framework::ThreadPool* pool, std::atomic<int>& sum, int cnt) {
   std::vector<std::future<void>> fs;
   for (int i = 0; i < cnt; ++i) {
-    auto f = pool->Run([&sum]() { sum.fetch_add(1); });
-    fs.push_back(std::move(f));
-  }
-  for (auto& f : fs) {
-    f.wait();
+    fs.push_back(framework::Async([&sum]() { sum.fetch_add(1); }));
   }
 }
 

paddle/inference/inference.cc

@@ -15,7 +15,6 @@ limitations under the License. */
 #include "inference.h"
 #include <fstream>
 #include "paddle/framework/executor.h"
-#include "paddle/framework/feed_fetch_method.h"
 #include "paddle/framework/init.h"
 #include "paddle/framework/scope.h"
 
@@ -154,7 +153,7 @@ void InferenceEngine::Execute(const std::vector<framework::LoDTensor>& feeds,
     LOG(FATAL) << "Please initialize the program_ and load_program_ first.";
   }
 
-  if (feeds.size() < feed_var_names_.size()) {
+  if (feeds.size() != feed_var_names_.size()) {
     LOG(FATAL) << "Please feed " << feed_var_names_.size() << " input Tensors.";
   }
 
@@ -165,19 +164,22 @@ void InferenceEngine::Execute(const std::vector<framework::LoDTensor>& feeds,
 
   executor->Run(*load_program_, scope, 0, true, true);
 
+  std::map<std::string, const framework::LoDTensor*> feed_targets;
+  std::map<std::string, framework::LoDTensor*> fetch_targets;
+
   // set_feed_variable
   for (size_t i = 0; i < feed_var_names_.size(); ++i) {
-    framework::SetFeedVariable(scope, feeds[i], "feed", i);
+    feed_targets[feed_var_names_[i]] = &feeds[i];
   }
 
-  executor->Run(*program_, scope, 0, true, true);
-
   // get_fetch_variable
   fetchs.resize(fetch_var_names_.size());
   for (size_t i = 0; i < fetch_var_names_.size(); ++i) {
-    fetchs[i] = framework::GetFetchVariable(*scope, "fetch", i);
+    fetch_targets[fetch_var_names_[i]] = &fetchs[i];
   }
 
+  executor->Run(*program_, scope, feed_targets, fetch_targets);
+
   delete place;
   delete scope;
   delete executor;

paddle/operators/activation_op.h

@@ -323,7 +323,7 @@ template <typename T>
 struct FloorFunctor : public BaseActivationFunctor<T> {
   template <typename Device, typename X, typename Out>
   void operator()(Device d, X x, Out out) const {
-    out.device(d) = x.ceil();
+    out.device(d) = x.floor();
   }
 };
 

paddle/operators/detail/grpc_client.cc

@@ -101,8 +101,8 @@ bool RPCClient::Wait() {
   if (req_count_ <= 0) {
     return true;
   }
-
-  std::vector<bool> a(req_count_);
+  const size_t kReqCnt = req_count_;
+  bool a[kReqCnt];
   std::vector<std::future<void>> waits(req_count_);
 
   for (int i = 0; i < req_count_; i++) {

paddle/operators/gru_op.cc

@@ -135,14 +135,14 @@ class GRUOpMaker : public framework::OpProtoAndCheckerMaker {
     AddComment(R"DOC(
 GRU Operator implements part calculations of the complete GRU as following:
 
-\f[
-update \ gate: u_t = actGate(xu_t + W_u * h_{t-1} + b_u) \\
-reset \ gate: r_t = actGate(xr_t + W_r * h_{t-1} + b_r)  \\
-output \ candidate: {h}_t = actNode(xc_t + W_c * dot(r_t, h_{t-1}) + b_c) \\
+$$
+update\_gate: u_t = actGate(xu_t + W_u * h_{t-1} + b_u) \\
+reset\_gate: r_t = actGate(xr_t + W_r * h_{t-1} + b_r)  \\
+output\_candidate: {h}_t = actNode(xc_t + W_c * dot(r_t, h_{t-1}) + b_c) \\
 output: h_t = dot((1 - u_t), h_{t-1}) + dot(u_t, {h}_t)
-\f]
+$$
 
-@note To implement the complete GRU, fully-connected operator must be used  
+@note To implement the complete GRU, fully-connected operator must be used
 before to feed xu, xr and xc as the Input of GRU operator.
 )DOC");
   }

paddle/operators/lstmp_op.cc

+/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License. */
+
+#include "paddle/operators/lstmp_op.h"
+
+namespace paddle {
+namespace operators {
+
+class LSTMPOp : public framework::OperatorWithKernel {
+ public:
+  using framework::OperatorWithKernel::OperatorWithKernel;
+
+  void InferShape(framework::InferShapeContext* ctx) const override {
+    PADDLE_ENFORCE(ctx->HasInput("Input"),
+                   "Input(Input) of LSTMP operator should not be null.");
+    PADDLE_ENFORCE(ctx->HasInput("Weight"),
+                   "Input(Weight) of LSTMP operator should not be null.");
+    PADDLE_ENFORCE(ctx->HasInput("ProjWeight"),
+                   "Input(ProjWeight) of LSTMP operator should not be null.");
+    PADDLE_ENFORCE(ctx->HasInput("Bias"),
+                   "Input(Bias) of LSTMP operator should not be null.");
+
+    PADDLE_ENFORCE(ctx->HasOutput("Projection"),
+                   "Output(Projection) of LSTMP operator should not be null.");
+    PADDLE_ENFORCE(ctx->HasOutput("Cell"),
+                   "Output(Cell) of LSTMP operator should not be null.");
+    PADDLE_ENFORCE(ctx->HasOutput("BatchGate"),
+                   "Output(BatchGate) of LSTMP operator should not be null.");
+    PADDLE_ENFORCE(ctx->HasOutput("BatchCellPreAct"),
+                   "Output(BatchCellPreAct) of LSTMP operator should not be "
+                   "null.");
+    PADDLE_ENFORCE(ctx->HasOutput("BatchHidden"),
+                   "Output(BatchHidden) of LSTMP operator should not be null.");
+
+    auto in_dims = ctx->GetInputDim("Input");
+    PADDLE_ENFORCE_EQ(in_dims.size(), 2,
+                      "Input(X)'s rank of LSTMP operator must be 2.");
+
+    int frame_size = in_dims[1] / 4;
+    auto w_dims = ctx->GetInputDim("Weight");
+    auto proj_dims = ctx->GetInputDim("ProjWeight");
+    PADDLE_ENFORCE_EQ(w_dims.size(), 2,
+                      "The rank of Input(Weight) should be 2.");
+    PADDLE_ENFORCE_EQ(w_dims[0], proj_dims[1],
+                      "The first dimension of Input(Weight) "
+                      "should be %d.",
+                      proj_dims[1]);
+    PADDLE_ENFORCE_EQ(w_dims[1], 4 * frame_size,
+                      "The second dimension of Input(Weight) "
+                      "should be 4 * %d.",
+                      frame_size);
+
+    PADDLE_ENFORCE_EQ(proj_dims.size(), 2,
+                      "The rank of Input(ProjWeight) should be 2.");
+    PADDLE_ENFORCE_EQ(proj_dims[0], frame_size,
+                      "The first dimension of Input(ProjWeight) "
+                      "should be %d.",
+                      frame_size);
+
+    if (ctx->HasInput("H0")) {
+      PADDLE_ENFORCE(ctx->HasInput("C0"),
+                     "Input(C0) of LSTMP operator should not be null after "
+                     "Input(H0) provided.");
+      auto h_dims = ctx->GetInputDim("H0");
+      auto c_dims = ctx->GetInputDim("C0");
+      PADDLE_ENFORCE(h_dims == c_dims,
+                     "The dimension of Input(H0) and Input(C0) "
+                     "should be the same.");
+      ctx->SetOutputDim("OrderedP0", {h_dims[0], proj_dims[1]});
+    }
+
+    auto b_dims = ctx->GetInputDim("Bias");
+    PADDLE_ENFORCE_EQ(b_dims.size(), 2, "The rank of Input(Bias) should be 2.");
+    PADDLE_ENFORCE_EQ(b_dims[0], 1,
+                      "The first dimension of Input(Bias) should be 1.");
+
+    if (ctx->Attrs().Get<bool>("use_peepholes")) {
+      PADDLE_ENFORCE_EQ(b_dims[1], 7 * frame_size,
+                        "The second dimension of Input(Bias) should be "
+                        "7 * %d if enable peepholes connection",
+                        frame_size);
+    } else {
+      PADDLE_ENFORCE_EQ(b_dims[1], 4 * frame_size,
+                        "The second dimension of Input(Bias) should be "
+                        "4 * %d if disable peepholes connection",
+                        frame_size);
+    }
+
+    framework::DDim out_dims({in_dims[0], frame_size});
+    framework::DDim proj_out_dims({in_dims[0], proj_dims[1]});
+    ctx->SetOutputDim("Projection", proj_out_dims);
+    ctx->SetOutputDim("Cell", out_dims);
+    ctx->SetOutputDim("BatchGate", in_dims);
+    ctx->SetOutputDim("BatchCellPreAct", out_dims);
+    ctx->SetOutputDim("BatchHidden", out_dims);
+    ctx->ShareLoD("Input", "Projection");
+    ctx->ShareLoD("Input", "Cell");
+  }
+
+ protected:
+  framework::OpKernelType GetExpectedKernelType(
+      const framework::ExecutionContext& ctx) const override {
+    return framework::OpKernelType(
+        framework::ToDataType(ctx.Input<framework::LoDTensor>("Input")->type()),
+        ctx.device_context());
+  }
+};
+
+class LSTMPOpMaker : public framework::OpProtoAndCheckerMaker {
+ public:
+  LSTMPOpMaker(OpProto* proto, OpAttrChecker* op_checker)
+      : OpProtoAndCheckerMaker(proto, op_checker) {
+    AddInput("Input",
+             "(LoDTensor) the input for sequence data, which supports "
+             "variable-time length input sequence. The underlying tensor in "
+             "this LoDTensor is a matrix with shape (T X 4D), where T is the "
+             "total time steps in this mini-batch, D is the hidden size.");
+    AddInput("H0",
+             "(Tensor, optional) the initial hidden state is an optional "
+             "input. This is a tensor with shape (N x D), where N is the "
+             "batch size and D is the hidden size.")
+        .AsDispensable();
+    AddInput("C0",
+             "(Tensor, optional) the initial cell state is an optional "
+             "input. This is a tensor with shape (N x D), where N is the "
+             "batch size. `C0` should not be null if `H0` provided.")
+        .AsDispensable();
+    AddInput("Weight",
+             "(Tensor) the learnable hidden-hidden weights."
+             " - The shape is (P x 4D), where P is the projection layer size "
+             "and  D is the hidden size."
+             " - Weight = {W_cr, W_ir, W_fr, W_or}");
+    AddInput("ProjWeight",
+             "(Tensor) the learnable weight of the projection layer."
+             " - The shape is (D x P), where P is the recurrent projection "
+             "layer size and  D is the hidden size."
+             " - ProjWeight = {W_rh}");
+    AddInput("Bias",
+             "(Tensor) the learnable biases, which contains two parts: "
+             "input-hidden biases and peephole connections weights if "
+             "setting `use_peepholes` to `True`. "
+             "1. `use_peepholes = False` "
+             " - The shape is (1 x 4D). "
+             " - Bias = {b_c, b_i, b_f, b_o}."
+             "2. `use_peepholes = True` "
+             " - The shape is (1 x 7D). "
+             " - Bias = {b_c, b_i, b_f, b_o, W_ic, W_fc, W_oc}.");
+    AddOutput("Projection",
+              "(LoDTensor) the projection of the hidden state of LSTMP "
+              "operator. The shape is (T x P), and LoD is the same with the "
+              "`Input`.");
+    AddOutput("Cell",
+              "(LoDTensor) the cell state of LSTMP operator. "
+              "The shape is (T x D), and lod is the same with the `Input`.");
+    AddOutput("BatchGate",
+              "(LoDTensor) This LoDTensor contains input gate, forget gate "
+              "and output gate after the activations. This LoDTensor has the "
+              "same shape as the reorganized input, which is also be called "
+              "batch input. The LoD size is 2. The first-level LoD is the "
+              "batch offsets and the second contains the indices, which "
+              "denotes the position of reorganized sequence in the raw input.")
+        .AsIntermediate();
+    AddOutput("BatchCellPreAct",
+              "(LoDTensor) the pre-activation cell state reorganized in batch. "
+              "This LoDTensor is obtained in the forward and used in the "
+              "backward.")
+        .AsIntermediate();
+    AddOutput("BatchHidden",
+              "(LoDTensor) the hidden state reorganized in batch. "
+              "This LoDTensor is obtained in the forward and used in the "
+              "backward.")
+        .AsIntermediate();
+    AddOutput("OrderedP0",
+              "(Tensor) the projection of the initial hidden state "
+              "H0. This is a tensor with shape (N x P), where N is the "
+              "batch size and P is the hidden size.")
+        .AsIntermediate();
+    AddAttr<bool>("use_peepholes",
+                  "(bool, defalut: True) "
+                  "whether to enable diagonal/peephole connections.")
+        .SetDefault(true);
+    AddAttr<bool>("is_reverse",
+                  "(bool, defalut: False) "
+                  "whether to compute reversed LSTMP.")
+        .SetDefault(false);
+    AddAttr<std::string>(
+        "gate_activation",
+        "(string, default: sigmoid)"
+        "The activation for input gate, forget gate and output "
+        "gate, `sigmoid` by default.")
+        .SetDefault("sigmoid")
+        .InEnum({"sigmoid", "tanh", "relu", "identity"});
+    AddAttr<std::string>("cell_activation",
+                         "(string, default: tanh)"
+                         "The activation for cell output, `tanh` by defalut.")
+        .SetDefault("tanh")
+        .InEnum({"sigmoid", "tanh", "relu", "identity"});
+    AddAttr<std::string>("candidate_activation",
+                         "(string, default: tanh)"
+                         "The activation for candidate hidden state, "
+                         "`tanh` by default.")
+        .SetDefault("tanh")
+        .InEnum({"sigmoid", "tanh", "relu", "identity"});
+    AddAttr<std::string>("proj_activation",
+                         "(string, default: tanh)"
+                         "The activation for projection output, "
+                         "`tanh` by defalut.")
+        .SetDefault("tanh")
+        .InEnum({"sigmoid", "tanh", "relu", "identity"});
+    AddComment(R"DOC(
+Long-Short Term Memory with recurrent Projection layer (LSTMP) Operator.
+
+LSTMP has a separate projection layer after the LSTM layer, projecting the 
+original hidden state to a lower-dimensional one, which is proposed to reduce 
+the number of total parameters and furthermore computational complexity for 
+the LSTM, espeacially for the case that the size of output units is relative 
+large (https://research.google.com/pubs/archive/43905.pdf). 
+
+The formula is as follows:
+
+$$
+i_t = \sigma(W_{ix}x_{t} + W_{ir}r_{t-1} + W_{ic}c_{t-1} + b_i) \\
+
+f_t = \sigma(W_{fx}x_{t} + W_{fr}r_{t-1} + W_{fc}c_{t-1} + b_f) \\
+
+\tilde{c_t} = act_g(W_{cx}x_t + W_{cr}r_{t-1} + b_c) \\
+
+o_t = \sigma(W_{ox}x_{t} + W_{or}r_{t-1} + W_{oc}c_t + b_o) \\
+
+c_t = f_t \odot c_{t-1} + i_t \odot \tilde{c_t} \\
+
+h_t = o_t \odot act_h(c_t) \\
+
+r_t = \overline{act_h}(W_{rh}h_t)
+$$
+
+where the W terms denote weight matrices (e.g. $W_{xi}$ is the matrix
+of weights from the input gate to the input), $W_{ic}, W_{fc}, W_{oc}$
+are diagonal weight matrices for peephole connections. In our implementation,
+we use vectors to reprenset these diagonal weight matrices. The b terms
+denote bias vectors ($b_i$ is the input gate bias vector), $\sigma$
+is the activation, such as logistic sigmoid function, and
+$i, f, o$ and $c$ are the input gate, forget gate, output gate,
+and cell activation vectors, respectively, all of which have the same size as
+the cell output activation vector $h$. Here $h$ is usually called the hidden 
+state and $r$ denotes its recurrent projection. And $\tilde{c_t}$ is also 
+called the candidate hidden state, whose computation is based on the current 
+input and previous hidden state.
+
+The $\odot$ is the element-wise product of the vectors. $act_g$ and $act_h$
+are the cell input and cell output activation functions and `tanh` is usually
+used for them. $\overline{act_h}$ is the activation function for the 
+projection output, usually using `identity` or same as $act_h$.
+
+Note that these $W_{xi}x_{t}, W_{xf}x_{t}, W_{xc}x_{t}, W_{xo}x_{t}$
+operations on the input $x_{t}$ are NOT included in this operator.
+Users can choose to use fully-connected operator before LSTMP operator.
+
+)DOC");
+  }
+};
+
+class LSTMPGradOp : public framework::OperatorWithKernel {
+ public:
+  using framework::OperatorWithKernel::OperatorWithKernel;
+
+  void InferShape(framework::InferShapeContext* ctx) const override {
+    PADDLE_ENFORCE(ctx->HasInput("Input"),
+                   "Input(Input) of LSTMP operator should not be null.");
+    PADDLE_ENFORCE(ctx->HasInput("Projection"),
+                   "Input(Projection) of LSTMP operator should not be null.");
+    PADDLE_ENFORCE(ctx->HasInput("Cell"),
+                   "Input(Cell) of LSTMP operator should not be null.");
+    PADDLE_ENFORCE(ctx->HasInput("Weight"),
+                   "Input(Weight) of LSTMP operator should not be null.");
+    PADDLE_ENFORCE(ctx->HasInput("ProjWeight"),
+                   "Input(ProjWeight) of LSTMP operator should not be null.");
+    PADDLE_ENFORCE(ctx->HasInput("Bias"),
+                   "Input(Bias) of LSTMP operator should not be null.");
+
+    PADDLE_ENFORCE(ctx->HasInput("BatchGate"),
+                   "Input(BatchGate) of LSTMP operator should not be null.");
+    PADDLE_ENFORCE(ctx->HasInput("BatchCellPreAct"),
+                   "Input(BatchGate) of LSTMP operator should not be null.");
+
+    auto SetOutGradDim = [&ctx](const std::string& name) {
+      auto g_name = framework::GradVarName(name);
+      if (ctx->HasOutput(g_name))
+        ctx->SetOutputDim(g_name, ctx->GetInputDim(name));
+    };
+
+    SetOutGradDim("Input");
+    SetOutGradDim("Weight");
+    SetOutGradDim("ProjWeight");
+    SetOutGradDim("Bias");
+    SetOutGradDim("H0");
+    SetOutGradDim("C0");
+  }
+
+ protected:
+  framework::OpKernelType GetExpectedKernelType(
+      const framework::ExecutionContext& ctx) const override {
+    return framework::OpKernelType(
+        framework::ToDataType(ctx.Input<framework::LoDTensor>("Input")->type()),
+        ctx.device_context());
+  }
+};
+
+}  // namespace operators
+}  // namespace paddle
+
+namespace ops = paddle::operators;
+REGISTER_OP(lstmp, ops::LSTMPOp, ops::LSTMPOpMaker, lstmp_grad,
+            ops::LSTMPGradOp);
+REGISTER_OP_CPU_KERNEL(
+    lstmp, ops::LSTMPKernel<paddle::platform::CPUDeviceContext, float>,
+    ops::LSTMPKernel<paddle::platform::CPUDeviceContext, double>);
+REGISTER_OP_CPU_KERNEL(
+    lstmp_grad, ops::LSTMPGradKernel<paddle::platform::CPUDeviceContext, float>,
+    ops::LSTMPGradKernel<paddle::platform::CPUDeviceContext, double>);

paddle/operators/lstmp_op.cu

+/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License. */
+
+#include "paddle/operators/lstmp_op.h"
+
+namespace ops = paddle::operators;
+REGISTER_OP_CUDA_KERNEL(
+    lstmp, ops::LSTMPKernel<paddle::platform::CUDADeviceContext, float>,
+    ops::LSTMPKernel<paddle::platform::CUDADeviceContext, double>);
+REGISTER_OP_CUDA_KERNEL(
+    lstmp_grad,
+    ops::LSTMPGradKernel<paddle::platform::CUDADeviceContext, float>,
+    ops::LSTMPGradKernel<paddle::platform::CUDADeviceContext, double>);

paddle/operators/multiplex_op.cc

@@ -119,7 +119,13 @@ REGISTER_OPERATOR(multiplex, ops::MultiplexOp, ops::MultiplexOpMaker,
 REGISTER_OPERATOR(multiplex_grad, ops::MultiplexGradOp);
 REGISTER_OP_CPU_KERNEL(
     multiplex,
-    ops::MultiplexCPUKernel<paddle::platform::CPUDeviceContext, float>);
+    ops::MultiplexCPUKernel<paddle::platform::CPUDeviceContext, float>,
+    ops::MultiplexCPUKernel<paddle::platform::CPUDeviceContext, double>,
+    ops::MultiplexCPUKernel<paddle::platform::CPUDeviceContext, int>,
+    ops::MultiplexCPUKernel<paddle::platform::CPUDeviceContext, int64_t>);
 REGISTER_OP_CPU_KERNEL(
     multiplex_grad,
-    ops::MultiplexGradCPUKernel<paddle::platform::CPUDeviceContext, float>);
+    ops::MultiplexGradCPUKernel<paddle::platform::CPUDeviceContext, float>,
+    ops::MultiplexGradCPUKernel<paddle::platform::CPUDeviceContext, double>,
+    ops::MultiplexGradCPUKernel<paddle::platform::CPUDeviceContext, int>,
+    ops::MultiplexGradCPUKernel<paddle::platform::CPUDeviceContext, int64_t>);

paddle/operators/multiplex_op.cu

@@ -90,7 +90,13 @@ namespace ops = paddle::operators;
 
 REGISTER_OP_CUDA_KERNEL(
     multiplex,
-    ops::MultiplexGPUKernel<paddle::platform::CUDADeviceContext, float>);
+    ops::MultiplexGPUKernel<paddle::platform::CUDADeviceContext, float>,
+    ops::MultiplexGPUKernel<paddle::platform::CUDADeviceContext, double>,
+    ops::MultiplexGPUKernel<paddle::platform::CUDADeviceContext, int>,
+    ops::MultiplexGPUKernel<paddle::platform::CUDADeviceContext, int64_t>);
 REGISTER_OP_CUDA_KERNEL(
     multiplex_grad,
-    ops::MultiplexGradGPUKernel<paddle::platform::CUDADeviceContext, float>);
+    ops::MultiplexGradGPUKernel<paddle::platform::CUDADeviceContext, float>,
+    ops::MultiplexGradGPUKernel<paddle::platform::CUDADeviceContext, double>,
+    ops::MultiplexGradGPUKernel<paddle::platform::CUDADeviceContext, int>,
+    ops::MultiplexGradGPUKernel<paddle::platform::CUDADeviceContext, int64_t>);

paddle/operators/one_hot_op.cc

+//   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#include "paddle/operators/one_hot_op.h"
+#include "paddle/framework/framework.pb.h"
+
+namespace paddle {
+namespace operators {
+
+class OneHotOp : public framework::OperatorWithKernel {
+ public:
+  using framework::OperatorWithKernel::OperatorWithKernel;
+  void InferShape(framework::InferShapeContext* ctx) const override {
+    PADDLE_ENFORCE(ctx->HasInput("X"),
+                   "Input(X) of OneHotOp should not be null.");
+    PADDLE_ENFORCE(ctx->HasOutput("Out"),
+                   "Output(Out) of OneHotOp should not be null.");
+
+    auto x_dims = ctx->GetInputDim("X");
+    PADDLE_ENFORCE_GE(x_dims.size(), 2,
+                      "Rank of Input(X) should be at least 2.");
+    PADDLE_ENFORCE_GE(x_dims[x_dims.size() - 1], 1U,
+                      "Last dimension of Input(X) should be 1.");
+
+    int depth = ctx->Attrs().Get<int>("depth");
+
+    PADDLE_ENFORCE_GT(depth, 0, "Should provide a positive depth (%d).", depth);
+
+    framework::DDim out_dims(x_dims);
+    out_dims[out_dims.size() - 1] = depth;
+    ctx->SetOutputDim("Out", out_dims);
+    ctx->ShareLoD("X", /* --> */ "Out");
+  }
+};
+
+class OneHotOpMaker : public framework::OpProtoAndCheckerMaker {
+ public:
+  OneHotOpMaker(OpProto* proto, OpAttrChecker* op_checker)
+      : OpProtoAndCheckerMaker(proto, op_checker) {
+    AddInput("X",
+             "(LoDTensor, LoDTensor<int>) Input variable with rank at least 2. "
+             "The last dimension of X should be 1. Each value of X is an index "
+             "to indicate the position.");
+    AddOutput("Out",
+              "(Tensor, Tensor<float>) Output tensor with same rank as X. "
+              "The tensor consists of one-hot representations of values in X.");
+    AddAttr<int>("depth",
+                 "A positive integer to specify the length of one-hot vector.");
+    AddAttr<int>("dtype",
+                 "An integer to specify the data type of one-hot "
+                 "vector. The default value is FP32.")
+        .SetDefault(paddle::framework::proto::DataType::FP32);
+    AddComment(R"DOC(
+One Hot Operator. This operator creates the one-hot representations for input
+index values. The following example will help to explain the function of this
+operator:
+
+X is a LoDTensor:
+  X.lod = [[0, 1, 4]]
+  X.shape = [4, 1]
+  X.data = [[1], [1], [3], [0]]
+
+set depth = 4
+
+Out is a LoDTensor:
+  Out.lod = [[0, 1, 4]]
+  Out.shape = [4, 4]
+  Out.data = [[0., 1., 0., 0.],
+              [0., 1., 0., 0.],
+              [0., 0., 0., 1.],
+              [1., 0., 0., 0.]]
+)DOC");
+  }
+};
+
+}  // namespace operators
+}  // namespace paddle
+
+namespace ops = paddle::operators;
+REGISTER_OPERATOR(one_hot, ops::OneHotOp, ops::OneHotOpMaker,
+                  paddle::framework::EmptyGradOpMaker);
+REGISTER_OP_CPU_KERNEL(
+    one_hot, ops::OneHotKernel<paddle::platform::CPUDeviceContext, int>,
+    ops::OneHotKernel<paddle::platform::CPUDeviceContext, int64_t>);

paddle/operators/one_hot_op.cu

+//   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#include "paddle/operators/one_hot_op.h"
+#include "paddle/platform/cuda_helper.h"
+#include "paddle/platform/gpu_info.h"
+
+namespace paddle {
+namespace operators {
+using platform::PADDLE_CUDA_NUM_THREADS;
+
+template <typename InT, typename OutT>
+__global__ void FillOutputKernel(const InT* p_in_data, OutT* p_out_data,
+                                 const int64_t numel, const int depth) {
+  int idx = blockIdx.x * blockDim.x + threadIdx.x;
+  if (idx < numel) {
+    *(p_out_data + (idx * depth) + p_in_data[idx]) = 1.0;
+  }
+}
+
+template <typename DeviceContext, typename InT>
+struct OneHotOpCUDAFunctor {
+  const framework::LoDTensor* in_;
+  framework::LoDTensor* out_;
+  const DeviceContext& ctx_;
+  int depth_;
+
+  OneHotOpCUDAFunctor(const framework::LoDTensor* in, framework::LoDTensor* out,
+                      int depth, const DeviceContext& ctx)
+      : in_(in), out_(out), depth_(depth), ctx_(ctx) {}
+
+  template <typename OutT>
+  void operator()() const {
+    auto* p_in_data = in_->data<InT>();
+    auto numel = in_->numel();
+    auto* p_out_data = out_->mutable_data<OutT>(ctx_.GetPlace());
+    auto stream = ctx_.stream();
+    math::set_constant(ctx_, out_, 0.0);
+
+    FillOutputKernel<<<(numel + PADDLE_CUDA_NUM_THREADS - 1) /
+                           PADDLE_CUDA_NUM_THREADS,
+                       PADDLE_CUDA_NUM_THREADS, 0, stream>>>(
+        p_in_data, p_out_data, numel, depth_);
+  }
+};
+
+using LoDTensor = framework::LoDTensor;
+template <typename DeviceContext, typename T>
+class OneHotCUDAKernel : public framework::OpKernel<T> {
+ public:
+  void Compute(const framework::ExecutionContext& context) const override {
+    auto* in = context.Input<LoDTensor>("X");
+    auto* out = context.Output<LoDTensor>("Out");
+    int depth = context.Attr<int>("depth");
+
+    framework::VisitDataType(
+        static_cast<framework::proto::DataType>(context.Attr<int>("dtype")),
+        OneHotOpCUDAFunctor<DeviceContext, T>(
+            in, out, depth, context.template device_context<DeviceContext>()));
+  }
+};
+
+}  // namespace operators
+}  // namespace paddle
+
+namespace ops = paddle::operators;
+REGISTER_OP_CUDA_KERNEL(
+    one_hot, ops::OneHotCUDAKernel<paddle::platform::CUDADeviceContext, int>,
+    ops::OneHotCUDAKernel<paddle::platform::CUDADeviceContext, int64_t>);

paddle/operators/one_hot_op.h

+//   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#pragma once
+#include "paddle/framework/op_registry.h"
+#include "paddle/operators/math/math_function.h"
+
+namespace paddle {
+namespace operators {
+
+template <typename DeviceContext, typename InT>
+struct OneHotOpFunctor {
+  const framework::LoDTensor* in_;
+  framework::LoDTensor* out_;
+  int depth_;
+  const DeviceContext& ctx_;
+
+  OneHotOpFunctor(const framework::LoDTensor* in, framework::LoDTensor* out,
+                  int depth, const DeviceContext& ctx)
+      : in_(in), out_(out), depth_(depth), ctx_(ctx) {}
+
+  template <typename OutT>
+  void operator()() const {
+    auto* p_in_data = in_->data<InT>();
+    auto numel = in_->numel();
+    auto* p_out_data = out_->mutable_data<OutT>(ctx_.GetPlace());
+    math::set_constant(ctx_, out_, 0.0);
+
+    for (int i = 0; i < numel; ++i) {
+      PADDLE_ENFORCE_GE(p_in_data[i], 0,
+                        "Illegal index value, should be at least 0.");
+      PADDLE_ENFORCE_LT(p_in_data[i], depth_,
+                        "Illegal index value, should be less than depth (%d).",
+                        depth_);
+      *(p_out_data + i * depth_ + p_in_data[i]) = 1.0;
+    }
+  }
+};
+
+using LoDTensor = framework::LoDTensor;
+template <typename DeviceContext, typename T>
+class OneHotKernel : public framework::OpKernel<T> {
+ public:
+  void Compute(const framework::ExecutionContext& context) const override {
+    auto* in = context.Input<LoDTensor>("X");
+    auto* out = context.Output<LoDTensor>("Out");
+    int depth = context.Attr<int>("depth");
+
+    framework::VisitDataType(
+        static_cast<framework::proto::DataType>(context.Attr<int>("dtype")),
+        OneHotOpFunctor<DeviceContext, T>(
+            in, out, depth, context.template device_context<DeviceContext>()));
+  }
+};
+
+}  // namespace operators
+}  // namespace paddle

paddle/operators/pool_op.h

@@ -139,10 +139,8 @@ class PoolGradKernel : public framework::OpKernel<T> {
     auto& dev_ctx = context.template device_context<DeviceContext>();
     if (in_x_grad) {
       in_x_grad->mutable_data<T>(context.GetPlace());
-      auto temp = framework::EigenVector<T>::Flatten(*in_x_grad);
-      temp.device(
-          *context.template device_context<DeviceContext>().eigen_device()) =
-          temp.constant(static_cast<T>(0));
+      paddle::operators::math::SetConstant<DeviceContext, T> set_constant;
+      set_constant(dev_ctx, in_x_grad, 0.0);
 
       switch (ksize.size()) {
         case 2: {

paddle/pybind/CMakeLists.txt

 if(WITH_PYTHON)
   cc_library(paddle_pybind SHARED
     SRCS pybind.cc exception.cc protobuf.cc const_value.cc
-    DEPS pybind python backward proto_desc paddle_memory executor prune init profiler
+    DEPS pybind python backward proto_desc paddle_memory executor prune init profiler feed_fetch_method
     ${GLOB_OP_LIB})
   if(NOT APPLE AND NOT ANDROID)
     target_link_libraries(paddle_pybind rt)

paddle/pybind/pybind.cc

@@ -424,7 +424,9 @@ All parameter, weight, gradient are variables in Paddle.
 
   py::class_<framework::Executor>(m, "Executor")
       .def(py::init<const platform::Place &>())
-      .def("run", &Executor::Run);
+      .def("run",
+           (void (Executor::*)(const ProgramDesc &, Scope *, int, bool, bool)) &
+               Executor::Run);
 
   m.def("unique_integer", UniqueIntegerGenerator);
   m.def("init_gflags", framework::InitGflags);

python/paddle/v2/fluid/distribute_transpiler.py

@@ -225,7 +225,7 @@ class DistributeTranspiler:
             if len(splited_vars) <= 1:
                 continue
             orig_var = program.global_block().vars[varname]
-            if orig_var == core.VarDesc.VarType.SELECTED_ROWS:
+            if orig_var.type == core.VarDesc.VarType.SELECTED_ROWS:
                 height_sections = []
                 for v in splited_vars:
                     height_sections.append(v.shape[0])
@@ -234,7 +234,7 @@ class DistributeTranspiler:
                     inputs={"X": orig_var},
                     outputs={"Out": splited_vars},
                     attrs={"height_sections": height_sections})
-            elif orig_var == core.VarDesc.VarType.LOD_TENSOR:
+            elif orig_var.type == core.VarDesc.VarType.LOD_TENSOR:
                 sections = []
                 for v in splited_vars:
                     sections.append(v.shape[0])

python/paddle/v2/fluid/layers/nn.py

@@ -63,6 +63,7 @@ __all__ = [
     'nce',
     'beam_search',
     'row_conv',
+    'multiplex',
 ]
 
 
@@ -2707,3 +2708,55 @@ def row_conv(input, future_context_size, param_attr=None, act=None):
                 'Filter': [filter_param]},
         outputs={'Out': [out]})
     return helper.append_activation(out)
+
+
+def multiplex(inputs, index):
+    """
+    **Multiplex Layer**
+
+    Referring to the given index variable, this layer selects rows from the
+    input variables to construct a multiplex variable. Assuming that there are
+    :math:`m` input variables and :math:`I_i` represents the i-th input
+    variable and :math:`i` is in [0, :math:`m`). All input variables are
+    tensors with same shape [:math:`d_0`, :math:`d_1`, ..., :math:`d_R`].
+    Please note that rank of the input tensor should be at least 2. Each input
+    variable will be treated as a 2-D matrix with shape [:math:`M`, :math:`N`]
+    where :math:`M` for :math:`d_0` and :math:`N` for :math:`d_1` * :math:`d_2`
+    * ... * :math:`d_R`. Let :math:`I_i[j]` be the j-th row of the i-th input
+    variable. The given index variable should be a 2-D tensor with shape
+    [:math:`M`, 1]. Let `ID[i]` be the i-th index value of the index variable.
+    Then the output variable will be a tensor with shape [:math:`d_0`,
+    :math:`d_1`, ..., :math:`d_R`]. If we treat the output tensor as a 2-D
+    matrix with shape [:math:`M`, :math:`N`] and let :math:`O[i]` be the i-th
+    row of the matrix, then `O[i]` is equal to :math:`I_{ID[i]}[i]`.
+
+    Args:
+       inputs (list): A list of variables to gather from. All variables have the
+                same shape and the rank is at least 2.
+       index (Variable): Tensor<int32>, index variable which is a 2-D tensor
+                with shape [M, 1] where M is the batch size.
+
+    Returns:
+        Variable: Multiplex variable gathered from input variables.
+
+    Examples:
+        .. code-block:: python
+
+            x1 = fluid.layers.data(name='x1', shape=[4], dtype='float32')
+            x2 = fluid.layers.data(name='x2', shape=[4], dtype='float32')
+            index = fluid.layers.data(name='index', shape=[1], dtype='int32')
+            out = fluid.layers.multiplex(inputs=[x1, x2], index=index)
+    """
+    helper = LayerHelper('multiplex', **locals())
+
+    if not isinstance(inputs, list) and len(inputs) < 2:
+        raise ValueError("inputs should be a list object and contains at least "
+                         "2 elements.")
+
+    out = helper.create_tmp_variable(inputs[0].dtype)
+    helper.append_op(
+        type='multiplex',
+        inputs={'X': inputs,
+                'Ids': index},
+        outputs={'Out': [out]})
+    return out

python/paddle/v2/fluid/tests/book/test_rnn_encoder_decoder.py

+#   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import numpy as np
+import paddle.v2 as paddle
+import paddle.v2.fluid as fluid
+import paddle.v2.fluid.core as core
+import paddle.v2.fluid.framework as framework
+import paddle.v2.fluid.layers as layers
+from paddle.v2.fluid.executor import Executor
+
+dict_size = 30000
+source_dict_dim = target_dict_dim = dict_size
+src_dict, trg_dict = paddle.dataset.wmt14.get_dict(dict_size)
+hidden_dim = 32
+embedding_dim = 16
+batch_size = 10
+max_length = 50
+topk_size = 50
+encoder_size = decoder_size = hidden_dim
+IS_SPARSE = True
+USE_PEEPHOLES = False
+
+
+def bi_lstm_encoder(input_seq, hidden_size):
+    input_forward_proj = fluid.layers.fc(input=input_seq,
+                                         size=hidden_size * 4,
+                                         bias_attr=True)
+    forward, _ = fluid.layers.dynamic_lstm(
+        input=input_forward_proj,
+        size=hidden_size * 4,
+        use_peepholes=USE_PEEPHOLES)
+    input_backward_proj = fluid.layers.fc(input=input_seq,
+                                          size=hidden_size * 4,
+                                          bias_attr=True)
+    backward, _ = fluid.layers.dynamic_lstm(
+        input=input_backward_proj,
+        size=hidden_size * 4,
+        is_reverse=True,
+        use_peepholes=USE_PEEPHOLES)
+
+    forward_last = fluid.layers.sequence_last_step(input=forward)
+    backward_first = fluid.layers.sequence_first_step(input=backward)
+
+    return forward_last, backward_first
+
+
+# FIXME(peterzhang2029): Replace this function with the lstm_unit_op.
+def lstm_step(x_t, hidden_t_prev, cell_t_prev, size):
+    def linear(inputs):
+        return fluid.layers.fc(input=inputs, size=size, bias_attr=True)
+
+    forget_gate = fluid.layers.sigmoid(x=linear([hidden_t_prev, x_t]))
+    input_gate = fluid.layers.sigmoid(x=linear([hidden_t_prev, x_t]))
+    output_gate = fluid.layers.sigmoid(x=linear([hidden_t_prev, x_t]))
+    cell_tilde = fluid.layers.tanh(x=linear([hidden_t_prev, x_t]))
+
+    cell_t = fluid.layers.sums(input=[
+        fluid.layers.elementwise_mul(
+            x=forget_gate, y=cell_t_prev), fluid.layers.elementwise_mul(
+                x=input_gate, y=cell_tilde)
+    ])
+
+    hidden_t = fluid.layers.elementwise_mul(
+        x=output_gate, y=fluid.layers.tanh(x=cell_t))
+
+    return hidden_t, cell_t
+
+
+def lstm_decoder_without_attention(target_embedding, decoder_boot, context,
+                                   decoder_size):
+    rnn = fluid.layers.DynamicRNN()
+
+    cell_init = fluid.layers.fill_constant_batch_size_like(
+        input=decoder_boot,
+        value=0.0,
+        shape=[-1, decoder_size],
+        dtype='float32')
+    cell_init.stop_gradient = False
+
+    with rnn.block():
+        current_word = rnn.step_input(target_embedding)
+        context = rnn.static_input(context)
+
+        hidden_mem = rnn.memory(init=decoder_boot, need_reorder=True)
+        cell_mem = rnn.memory(init=cell_init)
+        decoder_inputs = fluid.layers.concat(
+            input=[context, current_word], axis=1)
+        h, c = lstm_step(decoder_inputs, hidden_mem, cell_mem, decoder_size)
+        rnn.update_memory(hidden_mem, h)
+        rnn.update_memory(cell_mem, c)
+        out = fluid.layers.fc(input=h,
+                              size=target_dict_dim,
+                              bias_attr=True,
+                              act='softmax')
+        rnn.output(out)
+    return rnn()
+
+
+def seq_to_seq_net():
+    """Construct a seq2seq network."""
+
+    src_word_idx = fluid.layers.data(
+        name='source_sequence', shape=[1], dtype='int64', lod_level=1)
+
+    src_embedding = fluid.layers.embedding(
+        input=src_word_idx,
+        size=[source_dict_dim, embedding_dim],
+        dtype='float32')
+
+    src_forward_last, src_backward_first = bi_lstm_encoder(
+        input_seq=src_embedding, hidden_size=encoder_size)
+
+    encoded_vector = fluid.layers.concat(
+        input=[src_forward_last, src_backward_first], axis=1)
+
+    decoder_boot = fluid.layers.fc(input=src_backward_first,
+                                   size=decoder_size,
+                                   bias_attr=False,
+                                   act='tanh')
+
+    trg_word_idx = fluid.layers.data(
+        name='target_sequence', shape=[1], dtype='int64', lod_level=1)
+
+    trg_embedding = fluid.layers.embedding(
+        input=trg_word_idx,
+        size=[target_dict_dim, embedding_dim],
+        dtype='float32')
+
+    prediction = lstm_decoder_without_attention(trg_embedding, decoder_boot,
+                                                encoded_vector, decoder_size)
+    label = fluid.layers.data(
+        name='label_sequence', shape=[1], dtype='int64', lod_level=1)
+    cost = fluid.layers.cross_entropy(input=prediction, label=label)
+    avg_cost = fluid.layers.mean(x=cost)
+
+    return avg_cost
+
+
+def to_lodtensor(data, place):
+    seq_lens = [len(seq) for seq in data]
+    cur_len = 0
+    lod = [cur_len]
+    for l in seq_lens:
+        cur_len += l
+        lod.append(cur_len)
+    flattened_data = np.concatenate(data, axis=0).astype("int64")
+    flattened_data = flattened_data.reshape([len(flattened_data), 1])
+    res = core.LoDTensor()
+    res.set(flattened_data, place)
+    res.set_lod([lod])
+    return res
+
+
+def main():
+    avg_cost = seq_to_seq_net()
+
+    optimizer = fluid.optimizer.Adagrad(learning_rate=1e-4)
+    optimizer.minimize(avg_cost)
+
+    train_data = paddle.batch(
+        paddle.reader.shuffle(
+            paddle.dataset.wmt14.train(dict_size), buf_size=1000),
+        batch_size=batch_size)
+
+    place = core.CPUPlace()
+    exe = Executor(place)
+
+    exe.run(framework.default_startup_program())
+
+    batch_id = 0
+    for pass_id in xrange(2):
+        for data in train_data():
+            word_data = to_lodtensor(map(lambda x: x[0], data), place)
+            trg_word = to_lodtensor(map(lambda x: x[1], data), place)
+            trg_word_next = to_lodtensor(map(lambda x: x[2], data), place)
+            outs = exe.run(framework.default_main_program(),
+                           feed={
+                               'source_sequence': word_data,
+                               'target_sequence': trg_word,
+                               'label_sequence': trg_word_next
+                           },
+                           fetch_list=[avg_cost])
+            avg_cost_val = np.array(outs[0])
+            print('pass_id=' + str(pass_id) + ' batch=' + str(batch_id) +
+                  " avg_cost=" + str(avg_cost_val))
+            if batch_id > 3:
+                exit(0)
+            batch_id += 1
+
+
+if __name__ == '__main__':
+    main()

python/paddle/v2/fluid/tests/book_distribute/notest_dist_image_classification.py

-#Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved
+#   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserved.
 #
-#Licensed under the Apache License, Version 2.0 (the "License");
-#you may not use this file except in compliance with the License.
-#You may obtain a copy of the License at
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
 #
-#    http://www.apache.org/licenses/LICENSE-2.0
+#     http://www.apache.org/licenses/LICENSE-2.0
 #
-#Unless required by applicable law or agreed to in writing, software
-#distributed under the License is distributed on an "AS IS" BASIS,
-#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-#See the License for the specific language governing permissions and
-#limitations under the License.
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
 
 from __future__ import print_function
 
-import sys
-
 import paddle.v2 as paddle
 import paddle.v2.fluid as fluid
 import os
@@ -106,10 +104,10 @@ if len(sys.argv) >= 2:
     net_type = sys.argv[1]
 
 if net_type == "vgg":
-    print("train vgg net")
+    print("training vgg net")
     net = vgg16_bn_drop(images)
 elif net_type == "resnet":
-    print("train resnet")
+    print("training resnet")
     net = resnet_cifar10(images, 32)
 else:
     raise ValueError("%s network is not supported" % net_type)
@@ -129,6 +127,7 @@ train_reader = paddle.batch(
     batch_size=BATCH_SIZE)
 
 place = fluid.CPUPlace()
+feeder = fluid.DataFeeder(place=place, feed_list=[images, label])
 exe = fluid.Executor(place)
 
 t = fluid.DistributeTranspiler()
@@ -146,17 +145,14 @@ if training_role == "PSERVER":
     if not current_endpoint:
         print("need env SERVER_ENDPOINT")
         exit(1)
-    print("start pserver at:", current_endpoint)
     pserver_prog = t.get_pserver_program(current_endpoint)
     pserver_startup = t.get_startup_program(current_endpoint, pserver_prog)
     exe.run(pserver_startup)
     exe.run(pserver_prog)
-    print("pserver run end")
 elif training_role == "TRAINER":
-    print("start trainer")
     trainer_prog = t.get_trainer_program()
-    feeder = fluid.DataFeeder(place=place, feed_list=[images, label])
     exe.run(fluid.default_startup_program())
+
     for pass_id in range(PASS_NUM):
         accuracy.reset(exe)
         for data in train_reader():
@@ -164,9 +160,10 @@ elif training_role == "TRAINER":
                                 feed=feeder.feed(data),
                                 fetch_list=[avg_cost] + accuracy.metrics)
             pass_acc = accuracy.eval(exe)
-            print("loss:" + str(loss) + " acc:" + str(acc) + " pass_acc:" + str(
-                pass_acc))
-            # this model is slow, so if we can train two mini batch, we think it works properly.
+            print("pass_id:" + str(pass_id) + "loss:" + str(loss) + " pass_acc:"
+                  + str(pass_acc))
+            # this model is slow, so if we can train two mini batches,
+            # we think it works properly.
     print("trainer run end")
 else:
     print("environment var TRAINER_ROLE should be TRAINER os PSERVER")

python/paddle/v2/fluid/tests/book_distribute/notest_understand_sentiment_dynamic_lstm.py

+#   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import numpy as np
+import os
+import paddle.v2 as paddle
+import paddle.v2.fluid as fluid
+
+
+def stacked_lstm_net(data,
+                     label,
+                     input_dim,
+                     class_dim=2,
+                     emb_dim=128,
+                     hid_dim=512,
+                     stacked_num=3):
+    assert stacked_num % 2 == 1
+
+    emb = fluid.layers.embedding(input=data, size=[input_dim, emb_dim])
+    # add bias attr
+
+    # TODO(qijun) linear act
+    fc1 = fluid.layers.fc(input=emb, size=hid_dim)
+    lstm1, cell1 = fluid.layers.dynamic_lstm(input=fc1, size=hid_dim)
+
+    inputs = [fc1, lstm1]
+
+    for i in range(2, stacked_num + 1):
+        fc = fluid.layers.fc(input=inputs, size=hid_dim)
+        lstm, cell = fluid.layers.dynamic_lstm(
+            input=fc, size=hid_dim, is_reverse=(i % 2) == 0)
+        inputs = [fc, lstm]
+
+    fc_last = fluid.layers.sequence_pool(input=inputs[0], pool_type='max')
+    lstm_last = fluid.layers.sequence_pool(input=inputs[1], pool_type='max')
+
+    prediction = fluid.layers.fc(input=[fc_last, lstm_last],
+                                 size=class_dim,
+                                 act='softmax')
+    cost = fluid.layers.cross_entropy(input=prediction, label=label)
+    avg_cost = fluid.layers.mean(x=cost)
+    adam_optimizer = fluid.optimizer.Adam(learning_rate=0.002)
+    optimize_ops, params_grads = adam_optimizer.minimize(avg_cost)
+    accuracy = fluid.evaluator.Accuracy(input=prediction, label=label)
+    return avg_cost, accuracy, accuracy.metrics[0], optimize_ops, params_grads
+
+
+def to_lodtensor(data, place):
+    seq_lens = [len(seq) for seq in data]
+    cur_len = 0
+    lod = [cur_len]
+    for l in seq_lens:
+        cur_len += l
+        lod.append(cur_len)
+    flattened_data = np.concatenate(data, axis=0).astype("int64")
+    flattened_data = flattened_data.reshape([len(flattened_data), 1])
+    res = fluid.LoDTensor()
+    res.set(flattened_data, place)
+    res.set_lod([lod])
+    return res
+
+
+def main():
+    BATCH_SIZE = 100
+    PASS_NUM = 5
+
+    word_dict = paddle.dataset.imdb.word_dict()
+    print "loaded word dict successfully"
+    dict_dim = len(word_dict)
+    class_dim = 2
+
+    data = fluid.layers.data(
+        name="words", shape=[1], dtype="int64", lod_level=1)
+    label = fluid.layers.data(name="label", shape=[1], dtype="int64")
+    cost, accuracy, acc_out, optimize_ops, params_grads = stacked_lstm_net(
+        data, label, input_dim=dict_dim, class_dim=class_dim)
+
+    train_data = paddle.batch(
+        paddle.reader.shuffle(
+            paddle.dataset.imdb.train(word_dict), buf_size=1000),
+        batch_size=BATCH_SIZE)
+    place = fluid.CPUPlace()
+    exe = fluid.Executor(place)
+    feeder = fluid.DataFeeder(feed_list=[data, label], place=place)
+
+    t = fluid.DistributeTranspiler()
+    # all parameter server endpoints list for spliting parameters
+    pserver_endpoints = os.getenv("PSERVERS")
+    # server endpoint for current node
+    current_endpoint = os.getenv("SERVER_ENDPOINT")
+    # run as trainer or parameter server
+    training_role = os.getenv(
+        "TRAINING_ROLE", "TRAINER")  # get the training role: trainer/pserver
+    t.transpile(
+        optimize_ops, params_grads, pservers=pserver_endpoints, trainers=2)
+
+    if training_role == "PSERVER":
+        if not current_endpoint:
+            print("need env SERVER_ENDPOINT")
+            exit(1)
+        pserver_prog = t.get_pserver_program(current_endpoint)
+        pserver_startup = t.get_startup_program(current_endpoint, pserver_prog)
+        exe.run(pserver_startup)
+        exe.run(pserver_prog)
+    elif training_role == "TRAINER":
+        exe.run(fluid.default_startup_program())
+        trainer_prog = t.get_trainer_program()
+        for pass_id in xrange(PASS_NUM):
+            accuracy.reset(exe)
+            for data in train_data():
+                cost_val, acc_val = exe.run(trainer_prog,
+                                            feed=feeder.feed(data),
+                                            fetch_list=[cost, acc_out])
+                pass_acc = accuracy.eval(exe)
+                print("cost=" + str(cost_val) + " acc=" + str(acc_val) +
+                      " pass_acc=" + str(pass_acc))
+                if cost_val < 1.0 and acc_val > 0.8:
+                    exit(0)
+    else:
+        print("environment var TRAINER_ROLE should be TRAINER os PSERVER")
+
+
+if __name__ == '__main__':
+    main()

python/paddle/v2/fluid/tests/test_activation_op.py

@@ -186,8 +186,7 @@ class TestFloor(OpTest):
         self.op_type = "floor"
         x = np.random.uniform(-1, 1, [4, 4]).astype("float32")
         self.inputs = {'X': x}
-        # numpy floor need +1
-        self.outputs = {'Out': np.floor(self.inputs['X']) + 1.0}
+        self.outputs = {'Out': np.floor(self.inputs['X'])}
 
     def test_check_output(self):
         self.check_output()

python/paddle/v2/fluid/tests/test_layers.py

@@ -279,6 +279,16 @@ class TestBook(unittest.TestCase):
             self.assertIsNotNone(out)
         print(str(program))
 
+    def test_multiplex(self):
+        program = Program()
+        with program_guard(program):
+            x1 = layers.data(name='x1', shape=[4], dtype='float32')
+            x2 = layers.data(name='x2', shape=[4], dtype='float32')
+            index = layers.data(name='index', shape=[1], dtype='int32')
+            out = layers.multiplex(inputs=[x1, x2], index=index)
+            self.assertIsNotNone(out)
+        print(str(program))
+
 
 if __name__ == '__main__':
     unittest.main()

python/paddle/v2/fluid/tests/test_lstm_op.py

@@ -42,7 +42,7 @@ def relu(x):
     return np.maximum(x, 0)
 
 
-ACTVATION = {
+ACTIVATION = {
     'identity': identity,
     'sigmoid': sigmoid,
     'tanh': tanh,
@@ -158,8 +158,8 @@ class TestLstmOp(OpTest):
         w_b = b[:, 0:4 * self.D]
         w_c = b[:, 4 * self.D:] if self.use_peepholes else None
         h, c = lstm(x, self.lod, h0, c0, w, w_b, w_c, self.is_reverse,
-                    ACTVATION[self.act_gate], ACTVATION[self.act_cell],
-                    ACTVATION[self.act_cand])
+                    ACTIVATION[self.act_gate], ACTIVATION[self.act_cell],
+                    ACTIVATION[self.act_cand])
 
         self.inputs = {'Input': (x, self.lod), 'Weight': w}
 

python/paddle/v2/fluid/tests/test_lstmp_op.py

+#  Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+#
+#Licensed under the Apache License, Version 2.0 (the "License");
+#you may not use this file except in compliance with the License.
+#You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+#Unless required by applicable law or agreed to in writing, software
+#distributed under the License is distributed on an "AS IS" BASIS,
+#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#See the License for the specific language governing permissions and
+#limitations under the License.
+import unittest
+import numpy as np
+import test_lstm_op as LstmTest
+
+ACTIVATION = {
+    'identity': LstmTest.identity,
+    'sigmoid': LstmTest.sigmoid,
+    'tanh': LstmTest.tanh,
+    'relu': LstmTest.relu
+}
+
+
+# LSTM with recurrent projection Layer
+def lstmp(
+        input,  # T x 4D
+        lod,  # 1 x N
+        h0=None,  # N x D
+        c0=None,  # N x D
+        w_r=None,  # P x 4D
+        w_rh=None,  # D x P
+        w_b=None,  # 1 x 4D
+        w_c=None,  # 1 x 3D
+        is_reverse=False,
+        act_gate=None,
+        act_cell=None,
+        act_cand=None,
+        act_proj=None):
+    def _step(x, w_r, w_rh, w_c, r_pre, c_pre, act_gate, act_cell, act_cand,
+              act_proj):
+        g = np.dot(r_pre, w_r)  # 1 x 4D
+        g = g + x
+        g = np.reshape(g, (1, g.size))
+        c, g_i, g_f, g_o = np.split(g, 4, axis=1)
+        if w_c is None:
+            g_i = act_gate(g_i)  # 1 x D
+            g_f = act_gate(g_f)  # 1 x D
+        else:
+            w_ic, w_fc, _ = np.split(w_c, 3, axis=1)
+            g_i = act_gate(g_i + w_ic * c_pre)  # 1 x D
+            g_f = act_gate(g_f + w_fc * c_pre)  # 1 x D
+        c = g_f * c_pre + g_i * act_cand(c)  # 1 x D
+
+        if w_c is None:
+            g_o = act_gate(g_o)  # 1 x D
+        else:
+            _, _, w_oc = np.split(w_c, 3, axis=1)
+            g_o = act_gate(g_o + w_oc * c)  # 1 x D
+        h = g_o * act_cell(c)
+        # projection
+        r = np.dot(h, w_rh)
+        r = act_proj(r)
+        return r, c
+
+    def _reverse(x, lod):
+        y = np.zeros_like(x)
+        for i in range(len(lod) - 1):
+            b, e = lod[i], lod[i + 1]
+            y[b:e, :] = np.flip(x[b:e, :], 0)
+        return y
+
+    offset = lod[0]
+    batch_size = len(offset) - 1
+    # recurrent projection state
+    projection = []
+    cell = []
+    input = _reverse(input, offset) if is_reverse else input
+    if w_b is not None:
+        input = input + np.tile(w_b, (offset[-1], 1))
+    for i in range(batch_size):
+        # compute one sequence
+        seq_len = offset[i + 1] - offset[i]
+        x = input[offset[i]:offset[i + 1], :]
+        r_pre = np.dot(h0[i], w_rh)  # 1 x P
+        r_pre = act_proj(r_pre)
+        c_pre = c0[i]  # 1 x D
+        for j in range(seq_len):
+            # compute one step
+            r_pre, c_pre = _step(x[j], w_r, w_rh, w_c, r_pre, c_pre, act_gate,
+                                 act_cell, act_cand, act_proj)
+            projection.append(r_pre.flatten())
+            cell.append(c_pre.flatten())
+
+    projection = np.array(projection).astype('float64')
+    cell = np.array(cell).astype('float64')
+
+    projection = _reverse(projection, offset) if is_reverse else projection
+    cell = _reverse(cell, offset) if is_reverse else cell
+
+    assert projection.shape == (input.shape[0], w_r.shape[0])  # T x P
+    assert cell.shape == (input.shape[0], input.shape[1] / 4)  # T x D
+    return projection, cell
+
+
+class TestLstmpOp(LstmTest.TestLstmOp):
+    def reset_argument(self):
+        pass
+
+    def setUp(self):
+        self.set_argument()
+        # projection size
+        self.P = 10
+        self.act_proj = self.act_cell
+
+        self.reset_argument()
+        self.op_type = 'lstmp'
+
+        T = self.lod[0][-1]
+        N = len(self.lod[0]) - 1
+
+        x = np.random.normal(size=(T, 4 * self.D)).astype('float64')
+        if self.has_initial_state:
+            h0 = np.random.normal(size=(N, self.D)).astype('float64')
+            c0 = np.random.normal(size=(N, self.D)).astype('float64')
+        else:
+            h0 = np.zeros((N, self.D)).astype('float64')
+            c0 = np.zeros((N, self.D)).astype('float64')
+        w = np.random.normal(size=(self.P, 4 * self.D)).astype('float64')
+        if self.use_peepholes:
+            b = np.random.normal(size=(1, 7 * self.D)).astype('float64')
+        else:
+            b = np.random.normal(size=(1, 4 * self.D)).astype('float64')
+
+        w_b = b[:, 0:4 * self.D]
+        w_c = b[:, 4 * self.D:] if self.use_peepholes else None
+        w_rh = np.random.normal(size=(self.D, self.P)).astype('float64')
+        r, c = lstmp(x, self.lod, h0, c0, w, w_rh, w_b, w_c, self.is_reverse,
+                     ACTIVATION[self.act_gate], ACTIVATION[self.act_cell],
+                     ACTIVATION[self.act_cand], ACTIVATION[self.act_proj])
+
+        self.inputs = {'Input': (x, self.lod), 'Weight': w, 'ProjWeight': w_rh}
+
+        self.inputs['Bias'] = b
+
+        if self.has_initial_state:
+            self.inputs['H0'] = h0
+            self.inputs['C0'] = c0
+
+        self.outputs = {
+            'Projection': (r, self.lod),
+            'Cell': (c, self.lod),
+        }
+        self.attrs = {
+            'use_peepholes': self.use_peepholes,
+            'is_reverse': self.is_reverse,
+            'gate_activation': self.act_gate,
+            'cell_activation': self.act_cell,
+            'candidate_activation': self.act_cand,
+            'proj_activation': self.act_proj
+        }
+
+    def test_check_output(self):
+        self.check_output(atol=1e-8)
+
+    def test_check_grad(self):
+        # TODO(qingqing) remove folowing lines after the check_grad is refined.
+        N = len(self.lod[0]) - 1
+        self.outputs['OrderedP0'] = np.zeros((N, self.P)).astype('float64')
+        self.outputs['BatchGate'] = np.zeros((N, 4 * self.D)).astype('float64')
+        self.outputs['BatchHidden'] = np.zeros((N, self.D)).astype('float64')
+        self.outputs['BatchCellPreAct'] = np.zeros(
+            (N, self.D)).astype('float64')
+        self.check_grad(
+            ['Input', 'Weight', 'ProjWeight', 'Bias'], ['Projection'],
+            max_relative_error=1e-2)
+
+
+class TestLstmpOpHasInitial(TestLstmpOp):
+    def reset_argument(self):
+        self.has_initial_state = True
+
+    def test_check_grad(self):
+        # TODO(qingqing) remove folowing lines after the check_grad is refined.
+        N = len(self.lod[0]) - 1
+        self.outputs['OrderedP0'] = np.zeros((N, self.P)).astype('float64')
+        self.outputs['BatchGate'] = np.zeros((N, 4 * self.D)).astype('float64')
+        self.outputs['BatchHidden'] = np.zeros((N, self.D)).astype('float64')
+        self.outputs['BatchCellPreAct'] = np.zeros(
+            (N, self.D)).astype('float64')
+        self.check_grad(
+            ['Input', 'Weight', 'ProjWeight', 'Bias', 'H0', 'C0'],
+            ['Projection'],
+            max_relative_error=1e-2)
+
+    def test_check_grad_ingore_bias(self):
+        N = len(self.lod[0]) - 1
+        self.outputs['OrderedP0'] = np.zeros((N, self.P)).astype('float64')
+        self.outputs['BatchGate'] = np.zeros((N, 4 * self.D)).astype('float64')
+        self.outputs['BatchHidden'] = np.zeros((N, self.D)).astype('float64')
+        self.outputs['BatchCellPreAct'] = np.zeros(
+            (N, self.D)).astype('float64')
+        self.check_grad(
+            ['Input', 'ProjWeight', 'Weight'], ['Projection'],
+            max_relative_error=1e-2,
+            no_grad_set=set('Bias'))
+
+    def test_check_grad_ingore_weight(self):
+        N = len(self.lod[0]) - 1
+        self.outputs['OrderedP0'] = np.zeros((N, self.P)).astype('float64')
+        self.outputs['BatchGate'] = np.zeros((N, 4 * self.D)).astype('float64')
+        self.outputs['BatchHidden'] = np.zeros((N, self.D)).astype('float64')
+        self.outputs['BatchCellPreAct'] = np.zeros(
+            (N, self.D)).astype('float64')
+        self.check_grad(
+            ['Input', 'ProjWeight', 'Bias'], ['Projection'],
+            max_relative_error=1e-2,
+            no_grad_set=set('Weight'))
+
+    def test_check_grad_ingore_proj_weight(self):
+        N = len(self.lod[0]) - 1
+        self.outputs['OrderedP0'] = np.zeros((N, self.P)).astype('float64')
+        self.outputs['BatchGate'] = np.zeros((N, 4 * self.D)).astype('float64')
+        self.outputs['BatchHidden'] = np.zeros((N, self.D)).astype('float64')
+        self.outputs['BatchCellPreAct'] = np.zeros(
+            (N, self.D)).astype('float64')
+        self.check_grad(
+            ['Input', 'Weight', 'Bias'], ['Projection'],
+            max_relative_error=1e-2,
+            no_grad_set=set('ProjWeight'))
+
+    def test_check_grad_ingore_input(self):
+        N = len(self.lod[0]) - 1
+        self.outputs['OrderedP0'] = np.zeros((N, self.P)).astype('float64')
+        self.outputs['BatchGate'] = np.zeros((N, 4 * self.D)).astype('float64')
+        self.outputs['BatchHidden'] = np.zeros((N, self.D)).astype('float64')
+        self.outputs['BatchCellPreAct'] = np.zeros(
+            (N, self.D)).astype('float64')
+        self.check_grad(
+            ['Weight', 'ProjWeight', 'Bias'], ['Projection'],
+            max_relative_error=1e-2,
+            no_grad_set=set('Input'))
+
+    def test_check_grad_ingore_h0(self):
+        N = len(self.lod[0]) - 1
+        self.outputs['OrderedP0'] = np.zeros((N, self.P)).astype('float64')
+        self.outputs['BatchGate'] = np.zeros((N, 4 * self.D)).astype('float64')
+        self.outputs['BatchHidden'] = np.zeros((N, self.D)).astype('float64')
+        self.outputs['BatchCellPreAct'] = np.zeros(
+            (N, self.D)).astype('float64')
+        self.check_grad(
+            ['Input', 'Weight', 'ProjWeight', 'Bias', 'C0'], ['Projection'],
+            max_relative_error=1e-2,
+            no_grad_set=set('H0'))
+
+    def test_check_grad_ingore_c0(self):
+        N = len(self.lod[0]) - 1
+        self.outputs['OrderedP0'] = np.zeros((N, self.P)).astype('float64')
+        self.outputs['BatchGate'] = np.zeros((N, 4 * self.D)).astype('float64')
+        self.outputs['BatchHidden'] = np.zeros((N, self.D)).astype('float64')
+        self.outputs['BatchCellPreAct'] = np.zeros(
+            (N, self.D)).astype('float64')
+        self.check_grad(
+            ['Input', 'Weight', 'ProjWeight', 'Bias', 'H0'], ['Projection'],
+            max_relative_error=1e-2,
+            no_grad_set=set('C0'))
+
+
+class TestLstmpOpRerverse(TestLstmpOp):
+    def reset_argument(self):
+        self.is_reverse = True
+
+
+class TestLstmpOpNotUsePeepholes(TestLstmpOp):
+    def reset_argument(self):
+        self.use_peepholes = False
+
+
+class TestLstmpOpLinearProjection(TestLstmpOp):
+    def reset_argument(self):
+        self.act_proj = 'identity'
+
+
+if __name__ == '__main__':
+    unittest.main()

python/paddle/v2/fluid/tests/test_one_hot_op.py

+#   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import unittest
+import numpy as np
+import math
+from op_test import OpTest
+import paddle.v2.fluid as fluid
+import paddle.v2.fluid.core as core
+import paddle.v2.fluid.framework as framework
+from paddle.v2.fluid.framework import Program, program_guard
+
+
+class TestOneHotOp(OpTest):
+    def setUp(self):
+        self.op_type = 'one_hot'
+        depth = 10
+        dimension = 12
+        x_lod = [[0, 4, 5, 8, 11]]
+        x = [np.random.randint(0, depth - 1) for i in xrange(x_lod[0][-1])]
+        x = np.array(x).astype('int').reshape([x_lod[0][-1], 1])
+
+        out = np.zeros(shape=(np.product(x.shape[:-1]),
+                              depth)).astype('float32')
+
+        for i in xrange(np.product(x.shape)):
+            out[i, x[i]] = 1.0
+
+        self.inputs = {'X': (x, x_lod)}
+        self.attrs = {'depth': depth, 'dtype': int(core.DataType.FP32)}
+        self.outputs = {'Out': (out, x_lod)}
+
+    def test_check_output(self):
+        self.check_output()
+
+
+class TestOneHotOp_default_dtype(OpTest):
+    def setUp(self):
+        self.op_type = 'one_hot'
+        depth = 10
+        dimension = 12
+        x_lod = [[0, 4, 5, 8, 11]]
+        x = [np.random.randint(0, depth - 1) for i in xrange(x_lod[0][-1])]
+        x = np.array(x).astype('int').reshape([x_lod[0][-1], 1])
+
+        out = np.zeros(shape=(np.product(x.shape[:-1]),
+                              depth)).astype('float32')
+
+        for i in xrange(np.product(x.shape)):
+            out[i, x[i]] = 1.0
+
+        self.inputs = {'X': (x, x_lod)}
+        self.attrs = {'depth': depth}
+        self.outputs = {'Out': (out, x_lod)}
+
+    def test_check_output(self):
+        self.check_output()
+
+
+class TestOneHotOp_exception(OpTest):
+    def setUp(self):
+        self.op_type = 'one_hot'
+        self.depth = 10
+        self.place = core.CPUPlace()
+        self.dimension = 12
+        self.x = core.LoDTensor()
+        x_lod = [[0, 4, 5, 8, 11]]
+        data = [np.random.randint(11, 20) for i in xrange(x_lod[0][-1])]
+        data = np.array(data).astype('int').reshape([x_lod[0][-1], 1])
+        self.x.set(data, self.place)
+        self.x.set_lod(x_lod)
+
+    def test_check_output(self):
+        program = Program()
+        with program_guard(program):
+            x = fluid.layers.data(
+                name='x', shape=[self.dimension], dtype='float32', lod_level=1)
+            block = program.current_block()
+            one_hot_out = block.create_var(
+                name="one_hot_out",
+                type=core.VarDesc.VarType.LOD_TENSOR,
+                dtype='float32')
+            block.append_op(
+                type='one_hot',
+                inputs={'X': x},
+                attrs={'depth': self.depth},
+                outputs={'Out': one_hot_out})
+            exe = fluid.Executor(self.place)
+
+            def run():
+                exe.run(feed={'x': self.x},
+                        fetch_list=[one_hot_out],
+                        return_numpy=False)
+
+            self.assertRaises(core.EnforceNotMet, run)
+
+
+if __name__ == '__main__':
+    unittest.main()